rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgresql/src/backend/access/heap/vacuumlazy.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

3479 lines
114 KiB
C
Raw Normal View History

Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*-------------------------------------------------------------------------
*
* vacuumlazy.c
* Concurrent ("lazy") vacuuming.
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* The major space usage for vacuuming is storage for the array of dead TIDs
* that are to be removed from indexes. We want to ensure we can vacuum even
* the very largest relations with finite memory space usage. To do that, we
* set upper bounds on the number of TIDs we can keep track of at once.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
New autovacuum_work_mem parameter If autovacuum_work_mem is set, autovacuum workers now use this parameter in preference to maintenance_work_mem. Peter Geoghegan
2013-12-12 12:42:39 +01:00
* We are willing to use at most maintenance_work_mem (or perhaps
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* autovacuum_work_mem) memory space to keep track of dead TIDs. We initially
* allocate an array of TIDs of that size, with an upper limit that depends on
* table size (this limit ensures we don't allocate a huge area uselessly for
* vacuuming small tables). If the array threatens to overflow, we must call
* lazy_vacuum to vacuum indexes (and to vacuum the pages that we've pruned).
* This frees up the memory space dedicated to storing dead TIDs.
Code review for patch to avoid second scan when vacuuming index-less table: avoid invoking LockBufferForCleanup without need, put out the same log message we would have before, minor code beautification.
2006-09-13 19:47:08 +02:00
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* In practice VACUUM will often complete its initial pass over the target
* heap relation without ever running out of space to store TIDs. This means
* that there only needs to be one call to lazy_vacuum, after the initial pass
* completes.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
Update copyright for 2023 Backpatch-through: 11
2023-01-02 21:00:37 +01:00
* Portions Copyright (c) 1996-2023, PostgreSQL Global Development Group
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
* Portions Copyright (c) 1994, Regents of the University of California
*
*
* IDENTIFICATION
Move vacuumlazy.c into access/heap. It's heap table storage specific code that can't realistically be generalized into table AM agnostic code. Author: Andres Freund Discussion: https://postgr.es/m/20180703070645.wchpu5muyto5n647@alap3.anarazel.de
2019-01-15 21:06:19 +01:00
* src/backend/access/heap/vacuumlazy.c
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
*-------------------------------------------------------------------------
*/
#include "postgres.h"
Change planner to use the current true disk file size as its estimate of a relation's number of blocks, rather than the possibly-obsolete value in pg_class.relpages. Scale the value in pg_class.reltuples correspondingly to arrive at a hopefully more accurate number of rows. When pg_class contains 0/0, estimate a tuple width from the column datatypes and divide that into current file size to estimate number of rows. This improved methodology allows us to jettison the ancient hacks that put bogus default values into pg_class when a table is first created. Also, per a suggestion from Simon, make VACUUM (but not VACUUM FULL or ANALYZE) adjust the value it puts into pg_class.reltuples to try to represent the mean tuple density instead of the minimal density that actually prevails just after VACUUM. These changes alter the plans selected for certain regression tests, so update the expected files accordingly. (I removed join_1.out because it's not clear if it still applies; we can add back any variant versions as they are shown to be needed.)
2004-12-01 20:00:56 +01:00
#include <math.h>
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
#include "access/amapi.h"
Revert "Rename files and headers related to index AM" This follows multiple complains from Peter Geoghegan, Andres Freund and Alvaro Herrera that this issue ought to be dug more before actually happening, if it happens. Discussion: https://postgr.es/m/20191226144606.GA5659@alvherre.pgsql
2019-12-27 00:09:00 +01:00
#include "access/genam.h"
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
#include "access/heapam.h"
Split heapam_xlog.h from heapam.h The heapam XLog functions are used by other modules, not all of which are interested in the rest of the heapam API. With this, we let them get just the XLog stuff in which they are interested and not pollute them with unrelated includes. Also, since heapam.h no longer requires xlog.h, many files that do include heapam.h no longer get xlog.h automatically, including a few headers. This is useful because heapam.h is getting pulled in by execnodes.h, which is in turn included by a lot of files.
2012-08-29 01:02:00 +02:00
#include "access/heapam_xlog.h"
Split tuple struct defs from htup.h to htup_details.h This reduces unnecessary exposure of other headers through htup.h, which is very widely included by many files. I have chosen to move the function prototypes to the new file as well, because that means htup.h no longer needs to include tupdesc.h. In itself this doesn't have much effect in indirect inclusion of tupdesc.h throughout the tree, because it's also required by execnodes.h; but it's something to explore in the future, and it seemed best to do the htup.h change now while I'm busy with it.
2012-08-30 22:15:44 +02:00
#include "access/htup_details.h"
Improve concurrency of foreign key locking This patch introduces two additional lock modes for tuples: "SELECT FOR KEY SHARE" and "SELECT FOR NO KEY UPDATE". These don't block each other, in contrast with already existing "SELECT FOR SHARE" and "SELECT FOR UPDATE". UPDATE commands that do not modify the values stored in the columns that are part of the key of the tuple now grab a SELECT FOR NO KEY UPDATE lock on the tuple, allowing them to proceed concurrently with tuple locks of the FOR KEY SHARE variety. Foreign key triggers now use FOR KEY SHARE instead of FOR SHARE; this means the concurrency improvement applies to them, which is the whole point of this patch. The added tuple lock semantics require some rejiggering of the multixact module, so that the locking level that each transaction is holding can be stored alongside its Xid. Also, multixacts now need to persist across server restarts and crashes, because they can now represent not only tuple locks, but also tuple updates. This means we need more careful tracking of lifetime of pg_multixact SLRU files; since they now persist longer, we require more infrastructure to figure out when they can be removed. pg_upgrade also needs to be careful to copy pg_multixact files over from the old server to the new, or at least part of multixact.c state, depending on the versions of the old and new servers. Tuple time qualification rules (HeapTupleSatisfies routines) need to be careful not to consider tuples with the "is multi" infomask bit set as being only locked; they might need to look up MultiXact values (i.e. possibly do pg_multixact I/O) to find out the Xid that updated a tuple, whereas they previously were assured to only use information readily available from the tuple header. This is considered acceptable, because the extra I/O would involve cases that would previously cause some commands to block waiting for concurrent transactions to finish. Another important change is the fact that locking tuples that have previously been updated causes the future versions to be marked as locked, too; this is essential for correctness of foreign key checks. This causes additional WAL-logging, also (there was previously a single WAL record for a locked tuple; now there are as many as updated copies of the tuple there exist.) With all this in place, contention related to tuples being checked by foreign key rules should be much reduced. As a bonus, the old behavior that a subtransaction grabbing a stronger tuple lock than the parent (sub)transaction held on a given tuple and later aborting caused the weaker lock to be lost, has been fixed. Many new spec files were added for isolation tester framework, to ensure overall behavior is sane. There's probably room for several more tests. There were several reviewers of this patch; in particular, Noah Misch and Andres Freund spent considerable time in it. Original idea for the patch came from Simon Riggs, after a problem report by Joel Jacobson. Most code is from me, with contributions from Marti Raudsepp, Alexander Shulgin, Noah Misch and Andres Freund. This patch was discussed in several pgsql-hackers threads; the most important start at the following message-ids: AANLkTimo9XVcEzfiBR-ut3KVNDkjm2Vxh+t8kAmWjPuv@mail.gmail.com 1290721684-sup-3951@alvh.no-ip.org 1294953201-sup-2099@alvh.no-ip.org 1320343602-sup-2290@alvh.no-ip.org 1339690386-sup-8927@alvh.no-ip.org 4FE5FF020200002500048A3D@gw.wicourts.gov 4FEAB90A0200002500048B7D@gw.wicourts.gov
2013-01-23 16:04:59 +01:00
#include "access/multixact.h"
Allow include files to compile own their own. Strip unused include files out unused include files, and add needed includes to C files. The next step is to remove unused include files in C files.
2006-07-13 18:49:20 +02:00
#include "access/transam.h"
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
#include "access/visibilitymap.h"
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
#include "access/xact.h"
Move the backup-block logic from XLogInsert to a new file, xloginsert.c. xlog.c is huge, this makes it a little bit smaller, which is nice. Functions related to putting together the WAL record are in xloginsert.c, and the lower level stuff for managing WAL buffers and such are in xlog.c. Also move the definition of XLogRecord to a separate header file. This causes churn in the #includes of all the files that write WAL records, and redo routines, but it avoids pulling in xlog.h into most places. Reviewed by Michael Paquier, Alvaro Herrera, Andres Freund and Amit Kapila.
2014-11-06 12:52:08 +01:00
#include "access/xlog.h"
Remove xloginsert.h from xlog.h xlog.h is directly and indirectly #included in a lot of places. With this change, xloginsert.h is no longer unnecessarily included in the large number of them that don't need it. Author: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com> Discussion: https://postgr.es/m/CALj2ACVe-W+WM5P44N7eG9C2_FmaeM8Dq5aCnD3fHt0Ba=WR6w@mail.gmail.com
2022-01-30 16:25:24 +01:00
#include "access/xloginsert.h"
Add bound check before bsearch() for performance In the current lazy vacuum implementation, some index AMs such as btree indexes call lazy_tid_reaped() for each index tuple during ambulkdelete to check if the index tuple points to the (collected) garbage tuple. In that function, we simply call bsearch(), but we should be able to know the result without bsearch() if the index tuple points to the heap tuple that is out of range of the collected garbage tuples. Therefore, add a simple bound check before resorting to bsearch(). Testing has shown that this can give significant performance benefits. Author: Masahiko Sawada <masahiko.sawada@2ndquadrant.com> Discussion: https://www.postgresql.org/message-id/flat/CA+fd4k76j8jKzJzcx8UqEugvayaMSnQz0iLUt_XgBp-_-bd22A@mail.gmail.com
2021-03-10 15:19:37 +01:00
#include "catalog/index.h"
Rethink the way FSM truncation works. Instead of WAL-logging FSM truncations in FSM code, call FreeSpaceMapTruncateRel from smgr_redo. To make that cleaner from modularity point of view, move the WAL-logging one level up to RelationTruncate, and move RelationTruncate and all the related WAL-logging to new src/backend/catalog/storage.c file. Introduce new RelationCreateStorage and RelationDropStorage functions that are used instead of calling smgrcreate/smgrscheduleunlink directly. Move the pending rel deletion stuff from smgrcreate/smgrscheduleunlink to the new functions. This leaves smgr.c as a thin wrapper around md.c; all the transactional stuff is now in storage.c. This will make it easier to add new forks with similar truncation logic, like the visibility map.
2008-11-19 11:34:52 +01:00
#include "catalog/storage.h"
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
#include "commands/dbcommands.h"
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
#include "commands/progress.h"
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
#include "commands/vacuum.h"
Allow parallel vacuum to accumulate buffer usage. Commit 40d964ec99 allowed vacuum command to process indexes in parallel but forgot to accumulate the buffer usage stats of parallel workers. This allows leader backend to accumulate buffer usage stats of all the parallel workers. Reported-by: Julien Rouhaud Author: Sawada Masahiko Reviewed-by: Dilip Kumar, Amit Kapila and Julien Rouhaud Discussion: https://postgr.es/m/20200328151721.GB12854@nol
2020-04-02 04:34:58 +02:00
#include "executor/instrument.h"
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
#include "miscadmin.h"
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
#include "optimizer/paths.h"
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
#include "pgstat.h"
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
#include "portability/instr_time.h"
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
#include "postmaster/autovacuum.h"
Restructure some header files a bit, in particular heapam.h, by removing some unnecessary #include lines in it. Also, move some tuple routine prototypes and macros to htup.h, which allows removal of heapam.h inclusion from some .c files. For this to work, a new header file access/sysattr.h needed to be created, initially containing attribute numbers of system columns, for pg_dump usage. While at it, make contrib ltree, intarray and hstore header files more consistent with our header style.
2008-05-12 02:00:54 +02:00
#include "storage/bufmgr.h"
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
#include "storage/freespace.h"
Restructure some header files a bit, in particular heapam.h, by removing some unnecessary #include lines in it. Also, move some tuple routine prototypes and macros to htup.h, which allows removal of heapam.h inclusion from some .c files. For this to work, a new header file access/sysattr.h needed to be created, initially containing attribute numbers of system columns, for pg_dump usage. While at it, make contrib ltree, intarray and hstore header files more consistent with our header style.
2008-05-12 02:00:54 +02:00
#include "storage/lmgr.h"
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
#include "tcop/tcopprot.h"
Make VACUUM handle schema-qualified relation names properly.
2002-04-02 03:03:07 +02:00
#include "utils/lsyscache.h"
Prevent lazy_space_alloc from making requests that exceed MaxAllocSize, per report from Stefan Kaltenbrunner.
2006-03-04 20:09:09 +01:00
#include "utils/memutils.h"
Separate out the VacRUsage stuff as an independent module, in preparation for using it for other things besides VACUUM.
2005-10-04 00:52:26 +02:00
#include "utils/pg_rusage.h"
Move Timestamp/Interval typedefs and basic macros into datatype/timestamp.h. As per my recent proposal, this refactors things so that these typedefs and macros are available in a header that can be included in frontend-ish code. I also changed various headers that were undesirably including utils/timestamp.h to include datatype/timestamp.h instead. Unsurprisingly, this showed that half the system was getting utils/timestamp.h by way of xlog.h. No actual code changes here, just header refactoring.
2011-09-09 19:23:41 +02:00
#include "utils/timestamp.h"
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
* Space/time tradeoff parameters: do these need to be user-tunable?
*
* To consider truncating the relation, we want there to be at least
Reimplement free-space-map management as per recent discussions. Adjustable threshold is gone in favor of keeping track of total requested page storage and doling out proportional fractions to each relation (with a minimum amount per relation, and some quantization of the results to avoid thrashing with small changes in page counts). Provide special- case code for indexes so as not to waste space storing useless page free space counts. Restructure internal data storage to be a flat array instead of list-of-chunks; this may cost a little more work in data copying when reorganizing, but allows binary search to be used during lookup_fsm_page_entry().
2003-03-04 22:51:22 +01:00
* REL_TRUNCATE_MINIMUM or (relsize / REL_TRUNCATE_FRACTION) (whichever
* is less) potentially-freeable pages.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
Reimplement free-space-map management as per recent discussions. Adjustable threshold is gone in favor of keeping track of total requested page storage and doling out proportional fractions to each relation (with a minimum amount per relation, and some quantization of the results to avoid thrashing with small changes in page counts). Provide special- case code for indexes so as not to waste space storing useless page free space counts. Restructure internal data storage to be a flat array instead of list-of-chunks; this may cost a little more work in data copying when reorganizing, but allows binary search to be used during lookup_fsm_page_entry().
2003-03-04 22:51:22 +01:00
#define REL_TRUNCATE_MINIMUM 1000
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
#define REL_TRUNCATE_FRACTION 16
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* Timing parameters for truncate locking heuristics.
*
* These were not exposed as user tunable GUC values because it didn't seem
* that the potential for improvement was great enough to merit the cost of
* supporting them.
*/
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
#define VACUUM_TRUNCATE_LOCK_CHECK_INTERVAL 20 /* ms */
#define VACUUM_TRUNCATE_LOCK_WAIT_INTERVAL 50 /* ms */
#define VACUUM_TRUNCATE_LOCK_TIMEOUT 5000 /* ms */
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
/*
* Threshold that controls whether we bypass index vacuuming and heap
* vacuuming as an optimization
*/
#define BYPASS_THRESHOLD_PAGES 0.02 /* i.e. 2% of rel_pages */
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
/*
Use scanned_pages to decide when to failsafe check. Perform a failsafe check every time VACUUM's first heap scan scans a further FAILSAFE_EVERY_PAGES pages, rather than using an approach based on the number of physical blocks that our current blkno is from the blkno at the time of the previous failsafe check. That way VACUUM will perform a failsafe check every time it has scanned a uniform number of pages, without it mattering when or how VACUUM skipped pages using the visibility map. Sami Imseih, with changes to FAILSAFE_EVERY_PAGES comments added by me. Author: Sami Imseih <simseih@amazon.com> Reviewed-By: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/401CE010-4049-4B94-9961-0B610A5D254D%40amazon.com
2022-12-22 19:41:40 +01:00
* Perform a failsafe check each time we scan another 4GB of pages.
* (Note that this is deliberately kept to a power-of-two, usually 2^19.)
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
*/
Consider triggering VACUUM failsafe during scan. The wraparound failsafe mechanism added by commit 1e55e7d1 handled the one-pass strategy case (i.e. the "table has no indexes" case) by adding a dedicated failsafe check. This made up for the fact that the usual one-pass checks inside lazy_vacuum_all_indexes() cannot ever be reached during a one-pass strategy VACUUM. This approach failed to account for two-pass VACUUMs that opt out of index vacuuming up-front. The INDEX_CLEANUP off case in the only case that works like that. Fix this by performing a failsafe check every 4GB during the first scan of the heap, regardless of the details of the VACUUM. This eliminates the special case, and will make the failsafe trigger more reliably. Author: Peter Geoghegan <pg@bowt.ie> Reported-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/20210424002921.pb3t7h6frupdqnkp@alap3.anarazel.de
2021-05-25 02:14:02 +02:00
#define FAILSAFE_EVERY_PAGES \
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
((BlockNumber) (((uint64) 4 * 1024 * 1024 * 1024) / BLCKSZ))
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
/*
* When a table has no indexes, vacuum the FSM after every 8GB, approximately
* (it won't be exact because we only vacuum FSM after processing a heap page
* that has some removable tuples). When there are indexes, this is ignored,
* and we vacuum FSM after each index/heap cleaning pass.
*/
#define VACUUM_FSM_EVERY_PAGES \
((BlockNumber) (((uint64) 8 * 1024 * 1024 * 1024) / BLCKSZ))
Only skip pages marked as clean in the visibility map, if the last 32 pages were marked as clean as well. The idea is to avoid defeating OS readahead by skipping a page here and there, and also makes it less likely that we miss an opportunity to advance relfrozenxid, for the sake of only a few skipped pages.
2009-01-22 20:25:00 +01:00
/*
* Before we consider skipping a page that's marked as clean in
* visibility map, we must've seen at least this many clean pages.
*/
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
#define SKIP_PAGES_THRESHOLD ((BlockNumber) 32)
Only skip pages marked as clean in the visibility map, if the last 32 pages were marked as clean as well. The idea is to avoid defeating OS readahead by skipping a page here and there, and also makes it less likely that we miss an opportunity to advance relfrozenxid, for the sake of only a few skipped pages.
2009-01-22 20:25:00 +01:00
Prefetch blocks during lazy vacuum's truncation scan Vacuum truncation scan can be sped up on rotating media by prefetching blocks in forward direction. That makes the blocks already present in memory by the time they are needed, while also letting OS read-ahead kick in. The truncate scan has been measured to be five times faster than without this patch (that was on a slow disk, but it shouldn't hurt on fast disks.) Author: Álvaro Herrera, loosely based on a submission by Claudio Freire Discussion: https://postgr.es/m/CAGTBQpa6NFGO_6g_y_7zQx8L9GcHDSQKYdo1tGuh791z6PYgEg@mail.gmail.com
2017-01-23 16:55:18 +01:00
/*
* Size of the prefetch window for lazy vacuum backwards truncation scan.
* Needs to be a power of 2.
*/
#define PREFETCH_SIZE ((BlockNumber) 32)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
/*
* Macro to check if we are in a parallel vacuum. If true, we are in the
* parallel mode and the DSM segment is initialized.
*/
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
#define ParallelVacuumIsActive(vacrel) ((vacrel)->pvs != NULL)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Phases of vacuum during which we report error context. */
typedef enum
{
VACUUM_ERRCB_PHASE_UNKNOWN,
VACUUM_ERRCB_PHASE_SCAN_HEAP,
VACUUM_ERRCB_PHASE_VACUUM_INDEX,
VACUUM_ERRCB_PHASE_VACUUM_HEAP,
VACUUM_ERRCB_PHASE_INDEX_CLEANUP,
VACUUM_ERRCB_PHASE_TRUNCATE
} VacErrPhase;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
typedef struct LVRelState
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/* Target heap relation and its indexes */
Relation rel;
Relation *indrels;
int nindexes;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
/* Buffer access strategy and parallel vacuum state */
BufferAccessStrategy bstrategy;
ParallelVacuumState *pvs;
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
/* Aggressive VACUUM? (must set relfrozenxid >= FreezeLimit) */
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
bool aggressive;
/* Use visibility map to skip? (disabled by DISABLE_PAGE_SKIPPING) */
bool skipwithvm;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
/* Consider index vacuuming bypass optimization? */
bool consider_bypass_optimization;
/* Doing index vacuuming, index cleanup, rel truncation? */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
bool do_index_vacuuming;
bool do_index_cleanup;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
bool do_rel_truncate;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
/* VACUUM operation's cutoffs for freezing and pruning */
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
struct VacuumCutoffs cutoffs;
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
GlobalVisState *vistest;
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
/* Tracks oldest extant XID/MXID for setting relfrozenxid/relminmxid */
TransactionId NewRelfrozenXid;
MultiXactId NewRelminMxid;
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
bool skippedallvis;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/* Error reporting state */
vacuumlazy.c: Save get_database_name() in vacrel. This brings dbname strings in line with namespace and relation name strings. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkQ1TKU-DdNvnGeL870di3+CU1UTo-7nw7xFDpVE-XGjA@mail.gmail.com
2023-01-03 20:48:47 +01:00
char *dbname;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
char *relnamespace;
char *relname;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
char *indname; /* Current index name */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
BlockNumber blkno; /* used only for heap operations */
OffsetNumber offnum; /* used only for heap operations */
VacErrPhase phase;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
bool verbose; /* VACUUM VERBOSE? */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/*
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
* dead_items stores TIDs whose index tuples are deleted by index
* vacuuming. Each TID points to an LP_DEAD line pointer from a heap page
* that has been processed by lazy_scan_prune. Also needed by
* lazy_vacuum_heap_rel, which marks the same LP_DEAD line pointers as
* LP_UNUSED during second heap pass.
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
*/
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
VacDeadItems *dead_items; /* TIDs whose index tuples we'll delete */
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
BlockNumber rel_pages; /* total number of pages */
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
BlockNumber scanned_pages; /* # pages examined (not skipped via VM) */
vacuumlazy.c: Rename state field for consistency. Rename pages_removed to removed_pages, for consistency with nearby vacrel fields.
2022-01-29 02:41:09 +01:00
BlockNumber removed_pages; /* # pages removed by relation truncation */
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
BlockNumber frozen_pages; /* # pages with newly frozen tuples */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
BlockNumber lpdead_item_pages; /* # pages with LP_DEAD items */
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
BlockNumber missed_dead_pages; /* # pages with missed dead tuples */
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
BlockNumber nonempty_pages; /* actually, last nonempty page + 1 */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/* Statistics output by us, for table */
double new_rel_tuples; /* new estimated total # of tuples */
double new_live_tuples; /* new estimated total # of live tuples */
/* Statistics output by index AMs */
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
IndexBulkDeleteResult **indstats;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/* Instrumentation counters */
int num_index_scans;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Counters that follow are only for scanned_pages */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
int64 tuples_deleted; /* # deleted from table */
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
int64 tuples_frozen; /* # newly frozen */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
int64 lpdead_items; /* # deleted from indexes */
vacuumlazy.c: Remove obsolete num_tuples field. Commit 49c9d9fc unified VACUUM VERBOSE and autovacuum logging. It neglected to remove an old vacrel field that was only used by the old VACUUM VERBOSE, so remove it now. The previous num_tuples approach doesn't seem to have any real advantage over the approach VACUUM VERBOSE takes now (also the approach used by the autovacuum logging code), which is to show new_rel_tuples. new_rel_tuples is the possibly-estimated total number of tuples left in the table, whereas num_tuples meant the number of tuples encountered during the VACUUM operation, after pruning, without regard for tuples from pages skipped via the visibility map. In passing, reorder a related vacrel field for consistency.
2022-02-25 04:01:54 +01:00
int64 live_tuples; /* # live tuples remaining */
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
int64 recently_dead_tuples; /* # dead, but not yet removable */
int64 missed_dead_tuples; /* # removable, but not removed */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
} LVRelState;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
* State returned by lazy_scan_prune()
*/
typedef struct LVPagePruneState
{
Truncate line pointer array during heap pruning. Reclaim space from the line pointer array when heap pruning leaves behind a contiguous group of LP_UNUSED items at the end of the array. This happens during subsequent page defragmentation. Certain kinds of heap line pointer bloat are ameliorated by this new optimization. Follow-up work to commit 3c3b8a4b26, which taught VACUUM to truncate the line pointer array in about the same way during VACUUM's second pass over the heap. We now apply line pointer array truncation during both the first and the second pass over the heap made by VACUUM. We can also perform line pointer array truncation during opportunistic pruning. Matthias van de Meent, with small tweaks by me. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAEze2Wg36%2B4at2eWJNcYNiW2FJmht34x3YeX54ctUSs7kKoNcA%40mail.gmail.com
2022-04-08 00:42:12 +02:00
bool hastup; /* Page prevents rel truncation? */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
bool has_lpdead_items; /* includes existing LP_DEAD items */
/*
* State describes the proper VM bit states to set for the page following
* pruning and freezing. all_visible implies !has_lpdead_items, but don't
* trust all_frozen result unless all_visible is also set to true.
*/
bool all_visible; /* Every item visible to all? */
bool all_frozen; /* provided all_visible is also true */
TransactionId visibility_cutoff_xid; /* For recovery conflicts */
} LVPagePruneState;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
/* Struct for saving and restoring vacuum error information. */
typedef struct LVSavedErrInfo
{
BlockNumber blkno;
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
OffsetNumber offnum;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
VacErrPhase phase;
} LVSavedErrInfo;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/* non-export function prototypes */
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
static void lazy_scan_heap(LVRelState *vacrel);
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
static BlockNumber lazy_scan_skip(LVRelState *vacrel, Buffer *vmbuffer,
BlockNumber next_block,
bool *next_unskippable_allvis,
bool *skipping_current_range);
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
static bool lazy_scan_new_or_empty(LVRelState *vacrel, Buffer buf,
BlockNumber blkno, Page page,
bool sharelock, Buffer vmbuffer);
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
static void lazy_scan_prune(LVRelState *vacrel, Buffer buf,
BlockNumber blkno, Page page,
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
LVPagePruneState *prunestate);
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
static bool lazy_scan_noprune(LVRelState *vacrel, Buffer buf,
BlockNumber blkno, Page page,
bool *hastup, bool *recordfreespace);
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
static void lazy_vacuum(LVRelState *vacrel);
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
static bool lazy_vacuum_all_indexes(LVRelState *vacrel);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static void lazy_vacuum_heap_rel(LVRelState *vacrel);
static int lazy_vacuum_heap_page(LVRelState *vacrel, BlockNumber blkno,
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
Buffer buffer, int index, Buffer vmbuffer);
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
static bool lazy_check_wraparound_failsafe(LVRelState *vacrel);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static void lazy_cleanup_all_indexes(LVRelState *vacrel);
static IndexBulkDeleteResult *lazy_vacuum_one_index(Relation indrel,
IndexBulkDeleteResult *istat,
double reltuples,
LVRelState *vacrel);
static IndexBulkDeleteResult *lazy_cleanup_one_index(Relation indrel,
IndexBulkDeleteResult *istat,
double reltuples,
bool estimated_count,
LVRelState *vacrel);
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
static bool should_attempt_truncation(LVRelState *vacrel);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static void lazy_truncate_heap(LVRelState *vacrel);
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
static BlockNumber count_nondeletable_pages(LVRelState *vacrel,
bool *lock_waiter_detected);
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
static void dead_items_alloc(LVRelState *vacrel, int nworkers);
static void dead_items_cleanup(LVRelState *vacrel);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static bool heap_page_is_all_visible(LVRelState *vacrel, Buffer buf,
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
TransactionId *visibility_cutoff_xid, bool *all_frozen);
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
static void update_relstats_all_indexes(LVRelState *vacrel);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
static void vacuum_error_callback(void *arg);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static void update_vacuum_error_info(LVRelState *vacrel,
LVSavedErrInfo *saved_vacrel,
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
int phase, BlockNumber blkno,
OffsetNumber offnum);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static void restore_vacuum_error_info(LVRelState *vacrel,
const LVSavedErrInfo *saved_vacrel);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
Fix incorrect relation name in comment of vacuumlazy.c Author: Masahiko Sawada Discussion: https://postgr.es/m/CAD21AoBiOiapB7YGbWRfNZji3cs1gkEwv=uGLTemaZ9yNKK1DA@mail.gmail.com
2019-01-18 05:53:43 +01:00
* heap_vacuum_rel() -- perform VACUUM for one heap relation
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* This routine sets things up for and then calls lazy_scan_heap, where
* almost all work actually takes place. Finalizes everything after call
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
* returns by managing relation truncation and updating rel's pg_class
* entry. (Also updates pg_class entries for any indexes that need it.)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
* At entry, we have already established a transaction and opened
* and locked the relation.
*/
Fix up rickety handling of relation-truncation interlocks. Move rd_targblock, rd_fsm_nblocks, and rd_vm_nblocks from relcache to the smgr relation entries, so that they will get reset to InvalidBlockNumber whenever an smgr-level flush happens. Because we now send smgr invalidation messages immediately (not at end of transaction) when a relation truncation occurs, this ensures that other backends will reset their values before they next access the relation. We no longer need the unreliable assumption that a VACUUM that's doing a truncation will hold its AccessExclusive lock until commit --- in fact, we can intentionally release that lock as soon as we've completed the truncation. This patch therefore reverts (most of) Alvaro's patch of 2009-11-10, as well as my marginal hacking on it yesterday. We can also get rid of assorted no-longer-needed relcache flushes, which are far more expensive than an smgr flush because they kill a lot more state. In passing this patch fixes smgr_redo's failure to perform visibility-map truncation, and cleans up some rather dubious assumptions in freespace.c and visibilitymap.c about when rd_fsm_nblocks and rd_vm_nblocks can be out of date.
2010-02-09 22:43:30 +01:00
void
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
heap_vacuum_rel(Relation rel, VacuumParams *params,
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
BufferAccessStrategy bstrategy)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
LVRelState *vacrel;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
bool verbose,
vacuumlazy.c: Clean up variable declarations. Move some of the heap_vacuum_rel() instrumentation related variables to the scope where they're actually needed. Also reorder some of the variable declarations at the start of heap_vacuum_rel() so that related variables appear together.
2022-04-02 19:33:21 +02:00
instrument,
skipwithvm,
frozenxid_updated,
minmulti_updated;
BlockNumber orig_rel_pages,
new_rel_pages,
new_rel_allvisible;
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
PGRUsage ru0;
TimestampTz starttime = 0;
Fix multi-table VACUUM VERBOSE accounting. Per-backend global variables like VacuumPageHit are initialized once per VACUUM command. This was missed by commit 49c9d9fc, which unified VACUUM VERBOSE and autovacuum logging. As a result of that oversight, incorrect values were shown when multiple relations were processed by a single VACUUM VERBOSE command. Relations that happened to be processed later on would show "buffer usage:" values that incorrectly included buffer accesses made while processing earlier unrelated relations. The same accesses were counted multiple times. To fix, take initial values for the tracker variables at the start of heap_vacuum_rel(), and report delta values later on.
2022-04-16 00:48:39 +02:00
PgStat_Counter startreadtime = 0,
startwritetime = 0;
WalUsage startwalusage = pgWalUsage;
int64 StartPageHit = VacuumPageHit,
StartPageMiss = VacuumPageMiss,
StartPageDirty = VacuumPageDirty;
vacuumlazy.c: Clean up variable declarations. Move some of the heap_vacuum_rel() instrumentation related variables to the scope where they're actually needed. Also reorder some of the variable declarations at the start of heap_vacuum_rel() so that related variables appear together.
2022-04-02 19:33:21 +02:00
ErrorContextCallback errcallback;
char **indnames = NULL;
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
verbose = (params->options & VACOPT_VERBOSE) != 0;
instrument = (verbose || (IsAutoVacuumWorkerProcess() &&
params->log_min_duration >= 0));
if (instrument)
Make lazy_vacuum_rel call pg_rusage_init only if needed. do_analyze_rel already does it this way. Euler Taveira de Oliveira
2011-08-18 15:55:04 +02:00
{
pg_rusage_init(&ru0);
Improve logging of autovacuum I/O activity This adds some I/O stats to the logging of autovacuum (when the operation takes long enough that log_autovacuum_min_duration causes it to be logged), so that it is easier to tune. Notably, it adds buffer I/O counts (hits, misses, dirtied) and read and write rate. Authors: Greg Smith and Noah Misch
2011-11-25 16:10:46 +01:00
starttime = GetCurrentTimestamp();
Improve logging of auto-vacuum and auto-analyze When logging auto-vacuum and auto-analyze activity, include the I/O timing if track_io_timing is enabled. Also, for auto-analyze, add the read rate and the dirty rate, similar to how that information has historically been logged for auto-vacuum. Stephen Frost and Jakub Wartak Reviewed-By: Heikki Linnakangas, Tomas Vondra Discussion: https://www.postgresql.org/message-id/VI1PR0701MB69603A433348EDCF783C6ECBF6EF0%40VI1PR0701MB6960.eurprd07.prod.outlook.com
2021-03-16 19:46:48 +01:00
if (track_io_timing)
{
startreadtime = pgStatBlockReadTime;
startwritetime = pgStatBlockWriteTime;
}
Make lazy_vacuum_rel call pg_rusage_init only if needed. do_analyze_rel already does it this way. Euler Taveira de Oliveira
2011-08-18 15:55:04 +02:00
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Add a generic command progress reporting facility. Using this facility, any utility command can report the target relation upon which it is operating, if there is one, and up to 10 64-bit counters; the intent of this is that users should be able to figure out what a utility command is doing without having to resort to ugly hacks like attaching strace to a backend. As a demonstration, this adds very crude reporting to lazy vacuum; we just report the target relation and nothing else. A forthcoming patch will make VACUUM report a bunch of additional data that will make this much more interesting. But this gets the basic framework in place. Vinayak Pokale, Rahila Syed, Amit Langote, Robert Haas, reviewed by Kyotaro Horiguchi, Jim Nasby, Thom Brown, Masahiko Sawada, Fujii Masao, and Masanori Oyama.
2016-03-09 18:08:58 +01:00
pgstat_progress_start_command(PROGRESS_COMMAND_VACUUM,
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
RelationGetRelid(rel));
Make large sequential scans and VACUUMs work in a limited-size "ring" of buffers, rather than blowing out the whole shared-buffer arena. Aside from avoiding cache spoliation, this fixes the problem that VACUUM formerly tended to cause a WAL flush for every page it modified, because we had it hacked to use only a single buffer. Those flushes will now occur only once per ring-ful. The exact ring size, and the threshold for seqscans to switch into the ring usage pattern, remain under debate; but the infrastructure seems done. The key bit of infrastructure is a new optional BufferAccessStrategy object that can be passed to ReadBuffer operations; this replaces the former StrategyHintVacuum API. This patch also changes the buffer usage-count methodology a bit: we now advance usage_count when first pinning a buffer, rather than when last unpinning it. To preserve the behavior that a buffer's lifetime starts to decrease when it's released, the clock sweep code is modified to not decrement usage_count of pinned buffers. Work not done in this commit: teach GiST and GIN indexes to use the vacuum BufferAccessStrategy for vacuum-driven fetches. Original patch by Simon, reworked by Heikki and again by Tom.
2007-05-30 22:12:03 +02:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
/*
* Setup error traceback support for ereport() first. The idea is to set
* up an error context callback to display additional information on any
* error during a vacuum. During different phases of vacuum, we update
* the state so that the error context callback always display current
* information.
*
* Copy the names of heap rel into local memory for error reporting
* purposes, too. It isn't always safe to assume that we can get the name
* of each rel. It's convenient for code in lazy_scan_heap to always use
* these temp copies.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel = (LVRelState *) palloc0(sizeof(LVRelState));
vacuumlazy.c: Save get_database_name() in vacrel. This brings dbname strings in line with namespace and relation name strings. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkQ1TKU-DdNvnGeL870di3+CU1UTo-7nw7xFDpVE-XGjA@mail.gmail.com
2023-01-03 20:48:47 +01:00
vacrel->dbname = get_database_name(MyDatabaseId);
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
vacrel->relnamespace = get_namespace_name(RelationGetNamespace(rel));
vacrel->relname = pstrdup(RelationGetRelationName(rel));
vacrel->indname = NULL;
vacrel->phase = VACUUM_ERRCB_PHASE_UNKNOWN;
vacrel->verbose = verbose;
errcallback.callback = vacuum_error_callback;
errcallback.arg = vacrel;
errcallback.previous = error_context_stack;
error_context_stack = &errcallback;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
/* Set up high level stuff about rel and its indexes */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->rel = rel;
vac_open_indexes(vacrel->rel, RowExclusiveLock, &vacrel->nindexes,
&vacrel->indrels);
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
vacrel->bstrategy = bstrategy;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
if (instrument && vacrel->nindexes > 0)
{
/* Copy index names used by instrumentation (not error reporting) */
indnames = palloc(sizeof(char *) * vacrel->nindexes);
for (int i = 0; i < vacrel->nindexes; i++)
indnames[i] = pstrdup(RelationGetRelationName(vacrel->indrels[i]));
}
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
/*
* The index_cleanup param either disables index vacuuming and cleanup or
* forces it to go ahead when we would otherwise apply the index bypass
* optimization. The default is 'auto', which leaves the final decision
* up to lazy_vacuum().
*
* The truncate param allows user to avoid attempting relation truncation,
* though it can't force truncation to happen.
*/
Assert(params->index_cleanup != VACOPTVALUE_UNSPECIFIED);
Assert(params->truncate != VACOPTVALUE_UNSPECIFIED &&
params->truncate != VACOPTVALUE_AUTO);
Pre-beta mechanical code beautification. Run pgindent, pgperltidy, and reformat-dat-files. This set of diffs is a bit larger than typical. We've updated to pg_bsd_indent 2.1.2, which properly indents variable declarations that have multi-line initialization expressions (the continuation lines are now indented one tab stop). We've also updated to perltidy version 20230309 and changed some of its settings, which reduces its desire to add whitespace to lines to make assignments etc. line up. Going forward, that should make for fewer random-seeming changes to existing code. Discussion: https://postgr.es/m/20230428092545.qfb3y5wcu4cm75ur@alvherre.pgsql
2023-05-19 23:24:48 +02:00
Fix assertion failure in heap_vacuum_rel Commit 7d71d3dd08 changed resetting the VacuumFailsafeActive flag to an assertion since the flag is reset before starting vacuuming a relation. This however failed to take recursive calls of vacuum_rel() and vacuum of TOAST tables into consideration. Fix by reverting back to resettting the flag. Author: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Melanie Plageman <melanieplageman@gmail.com> Reported-by: John Naylor <john.naylor@enterprisedb.com> Discussion: https://postgr.es/m/CAFBsxsFz=GqaG5Ens5aNgVYoV2Y+pfMUijX0ku+CCkWfALwiqg@mail.gmail.com
2023-04-28 10:30:05 +02:00
/*
* While VacuumFailSafeActive is reset to false before calling this, we
* still need to reset it here due to recursive calls.
*/
VacuumFailsafeActive = false;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
vacrel->consider_bypass_optimization = true;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
vacrel->do_index_vacuuming = true;
vacrel->do_index_cleanup = true;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
vacrel->do_rel_truncate = (params->truncate != VACOPTVALUE_DISABLED);
if (params->index_cleanup == VACOPTVALUE_DISABLED)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
/* Force disable index vacuuming up-front */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
vacrel->do_index_vacuuming = false;
vacrel->do_index_cleanup = false;
}
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
else if (params->index_cleanup == VACOPTVALUE_ENABLED)
{
/* Force index vacuuming. Note that failsafe can still bypass. */
vacrel->consider_bypass_optimization = false;
}
else
{
/* Default/auto, make all decisions dynamically */
Assert(params->index_cleanup == VACOPTVALUE_AUTO);
}
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
/* Initialize page counters explicitly (be tidy) */
vacrel->scanned_pages = 0;
vacrel->removed_pages = 0;
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
vacrel->frozen_pages = 0;
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
vacrel->lpdead_item_pages = 0;
vacrel->missed_dead_pages = 0;
vacrel->nonempty_pages = 0;
/* dead_items_alloc allocates vacrel->dead_items later on */
/* Allocate/initialize output statistics state */
vacrel->new_rel_tuples = 0;
vacrel->new_live_tuples = 0;
vacrel->indstats = (IndexBulkDeleteResult **)
palloc0(vacrel->nindexes * sizeof(IndexBulkDeleteResult *));
/* Initialize remaining counters (be tidy) */
vacrel->num_index_scans = 0;
vacrel->tuples_deleted = 0;
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
vacrel->tuples_frozen = 0;
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
vacrel->lpdead_items = 0;
vacrel->live_tuples = 0;
vacrel->recently_dead_tuples = 0;
vacrel->missed_dead_tuples = 0;
/*
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
* Get cutoffs that determine which deleted tuples are considered DEAD,
* not just RECENTLY_DEAD, and which XIDs/MXIDs to freeze. Then determine
* the extent of the blocks that we'll scan in lazy_scan_heap. It has to
* happen in this order to ensure that the OldestXmin cutoff field works
* as an upper bound on the XIDs stored in the pages we'll actually scan
* (NewRelfrozenXid tracking must never be allowed to miss unfrozen XIDs).
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
*
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
* Next acquire vistest, a related cutoff that's used in heap_page_prune.
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
* We expect vistest will always make heap_page_prune remove any deleted
* tuple whose xmax is < OldestXmin. lazy_scan_prune must never become
* confused about whether a tuple should be frozen or removed. (In the
* future we might want to teach lazy_scan_prune to recompute vistest from
* time to time, to increase the number of dead tuples it can prune away.)
*/
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
vacrel->aggressive = vacuum_get_cutoffs(rel, params, &vacrel->cutoffs);
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
vacrel->rel_pages = orig_rel_pages = RelationGetNumberOfBlocks(rel);
vacrel->vistest = GlobalVisTestFor(rel);
vacuumlazy.c: MultiXactIds are MXIDs, not XMIDs. Oversights in commits 0b018fab and f3c15cbe.
2022-04-21 03:29:02 +02:00
/* Initialize state used to track oldest extant XID/MXID */
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
vacrel->NewRelfrozenXid = vacrel->cutoffs.OldestXmin;
vacrel->NewRelminMxid = vacrel->cutoffs.OldestMxact;
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
vacrel->skippedallvis = false;
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
skipwithvm = true;
if (params->options & VACOPT_DISABLE_PAGE_SKIPPING)
{
/*
* Force aggressive mode, and disable skipping blocks using the
* visibility map (even those set all-frozen)
*/
vacrel->aggressive = true;
skipwithvm = false;
}
vacrel->skipwithvm = skipwithvm;
if (verbose)
{
if (vacrel->aggressive)
ereport(INFO,
(errmsg("aggressively vacuuming \"%s.%s.%s\"",
vacuumlazy.c: Save get_database_name() in vacrel. This brings dbname strings in line with namespace and relation name strings. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkQ1TKU-DdNvnGeL870di3+CU1UTo-7nw7xFDpVE-XGjA@mail.gmail.com
2023-01-03 20:48:47 +01:00
vacrel->dbname, vacrel->relnamespace,
vacrel->relname)));
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
else
ereport(INFO,
(errmsg("vacuuming \"%s.%s.%s\"",
vacuumlazy.c: Save get_database_name() in vacrel. This brings dbname strings in line with namespace and relation name strings. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkQ1TKU-DdNvnGeL870di3+CU1UTo-7nw7xFDpVE-XGjA@mail.gmail.com
2023-01-03 20:48:47 +01:00
vacrel->dbname, vacrel->relnamespace,
vacrel->relname)));
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
/*
* Allocate dead_items array memory using dead_items_alloc. This handles
* parallel VACUUM initialization as part of allocating shared memory
* space used for dead_items. (But do a failsafe precheck first, to
* ensure that parallel VACUUM won't be attempted at all when relfrozenxid
* is already dangerously old.)
*/
lazy_check_wraparound_failsafe(vacrel);
dead_items_alloc(vacrel, params->nworkers);
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
/*
* Call lazy_scan_heap to perform all required heap pruning, index
* vacuuming, and heap vacuuming (plus related processing)
*/
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
lazy_scan_heap(vacrel);
/*
* Free resources managed by dead_items_alloc. This ends parallel mode in
* passing when necessary.
*/
dead_items_cleanup(vacrel);
Assert(!IsInParallelMode());
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Don't update relfrozenxid if any pages were skipped. Vacuum recognizes that it can update relfrozenxid by checking whether it has processed all pages of a relation. Unfortunately it performed that check after truncating the dead pages at the end of the relation, and used the new number of pages to decide whether all pages have been scanned. If the new number of pages happened to be smaller or equal to the number of pages scanned, it incorrectly decided that all pages were scanned. This can lead to relfrozenxid being updated, even though some pages were skipped that still contain old XIDs. That can lead to data loss due to xid wraparounds with some rows suddenly missing. This likely has escaped notice so far because it takes a large number (~2^31) of xids being used to see the effect, while a full-table vacuum before that would fix the issue. The incorrect logic was introduced by commit b4b6923e03f4d29636a94f6f4cc2f5cf6298b8c8. Backpatch this fix down to 8.4, like that commit. Andres Freund, with some modifications by me.
2013-11-27 12:10:16 +01:00
/*
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
* Update pg_class entries for each of rel's indexes where appropriate.
*
* Unlike the later update to rel's pg_class entry, this is not critical.
* Maintains relpages/reltuples statistics used by the planner only.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
if (vacrel->do_index_cleanup)
update_relstats_all_indexes(vacrel);
/* Done with rel's indexes */
vac_close_indexes(vacrel->nindexes, vacrel->indrels, NoLock);
/* Optionally truncate rel */
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
if (should_attempt_truncation(vacrel))
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_truncate_heap(vacrel);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Pop the error context stack */
error_context_stack = errcallback.previous;
Fix up rickety handling of relation-truncation interlocks. Move rd_targblock, rd_fsm_nblocks, and rd_vm_nblocks from relcache to the smgr relation entries, so that they will get reset to InvalidBlockNumber whenever an smgr-level flush happens. Because we now send smgr invalidation messages immediately (not at end of transaction) when a relation truncation occurs, this ensures that other backends will reset their values before they next access the relation. We no longer need the unreliable assumption that a VACUUM that's doing a truncation will hold its AccessExclusive lock until commit --- in fact, we can intentionally release that lock as soon as we've completed the truncation. This patch therefore reverts (most of) Alvaro's patch of 2009-11-10, as well as my marginal hacking on it yesterday. We can also get rid of assorted no-longer-needed relcache flushes, which are far more expensive than an smgr flush because they kill a lot more state. In passing this patch fixes smgr_redo's failure to perform visibility-map truncation, and cleans up some rather dubious assumptions in freespace.c and visibilitymap.c about when rd_fsm_nblocks and rd_vm_nblocks can be out of date.
2010-02-09 22:43:30 +01:00
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
/* Report that we are now doing final cleanup */
pgstat_progress_update_param(PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_PHASE_FINAL_CLEANUP);
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
/*
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
* Prepare to update rel's pg_class entry.
Fix a missed case in code for "moving average" estimate of reltuples. It is possible for VACUUM to scan no pages at all, if the visibility map shows that all pages are all-visible. In this situation VACUUM has no new information to report about the relation's tuple density, so it wasn't changing pg_class.reltuples ... but it updated pg_class.relpages anyway. That's wrong in general, since there is no evidence to justify changing the density ratio reltuples/relpages, but it's particularly bad if the previous state was relpages=reltuples=0, which means "unknown tuple density". We just replaced "unknown" with "zero". ANALYZE would eventually recover from this, but it could take a lot of repetitions of ANALYZE to do so if the relation size is much larger than the maximum number of pages ANALYZE will scan, because of the moving-average behavior introduced by commit b4b6923e03f4d29636a94f6f4cc2f5cf6298b8c8. The only known situation where we could have relpages=reltuples=0 and yet the visibility map asserts everything's visible is immediately following a pg_upgrade. It might be advisable for pg_upgrade to try to preserve the relpages/reltuples statistics; but in any case this code is wrong on its own terms, so fix it. Per report from Sergey Koposov. Back-patch to 8.4, where the visibility map was introduced, same as the previous change.
2011-08-30 20:49:45 +02:00
*
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
* Aggressive VACUUMs must always be able to advance relfrozenxid to a
* value >= FreezeLimit, and relminmxid to a value >= MultiXactCutoff.
* Non-aggressive VACUUMs may advance them by any amount, or not at all.
*/
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
Assert(vacrel->NewRelfrozenXid == vacrel->cutoffs.OldestXmin ||
TransactionIdPrecedesOrEquals(vacrel->aggressive ? vacrel->cutoffs.FreezeLimit :
vacrel->cutoffs.relfrozenxid,
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
vacrel->NewRelfrozenXid));
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
Assert(vacrel->NewRelminMxid == vacrel->cutoffs.OldestMxact ||
MultiXactIdPrecedesOrEquals(vacrel->aggressive ? vacrel->cutoffs.MultiXactCutoff :
vacrel->cutoffs.relminmxid,
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
vacrel->NewRelminMxid));
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
if (vacrel->skippedallvis)
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
{
/*
* Must keep original relfrozenxid in a non-aggressive VACUUM that
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
* chose to skip an all-visible page range. The state that tracks new
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
* values will have missed unfrozen XIDs from the pages we skipped.
*/
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
Assert(!vacrel->aggressive);
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
vacrel->NewRelfrozenXid = InvalidTransactionId;
vacrel->NewRelminMxid = InvalidMultiXactId;
}
/*
Redefine pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Historically, we've considered the state with relpages and reltuples both zero as indicating that we do not know the table's tuple density. This is problematic because it's impossible to distinguish "never yet vacuumed" from "vacuumed and seen to be empty". In particular, a user cannot use VACUUM or ANALYZE to override the planner's normal heuristic that an empty table should not be believed to be empty because it is probably about to get populated. That heuristic is a good safety measure, so I don't care to abandon it, but there should be a way to override it if the table is indeed intended to stay empty. Hence, represent the initial state of ignorance by setting reltuples to -1 (relpages is still set to zero), and apply the minimum-ten-pages heuristic only when reltuples is still -1. If the table is empty, VACUUM or ANALYZE (but not CREATE INDEX) will override that to reltuples = relpages = 0, and then we'll plan on that basis. This requires a bunch of fiddly little changes, but we can get rid of some ugly kluges that were formerly needed to maintain the old definition. One notable point is that FDWs' GetForeignRelSize methods will see baserel->tuples = -1 when no ANALYZE has been done on the foreign table. That seems like a net improvement, since those methods were formerly also in the dark about what baserel->tuples = 0 really meant. Still, it is an API change. I bumped catversion because code predating this change would get confused by seeing reltuples = -1. Discussion: https://postgr.es/m/F02298E0-6EF4-49A1-BCB6-C484794D9ACC@thebuild.com
2020-08-30 18:21:51 +02:00
* For safety, clamp relallvisible to be not more than what we're setting
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
* pg_class.relpages to
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
*/
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
new_rel_pages = vacrel->rel_pages; /* After possible rel truncation */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
visibilitymap_count(rel, &new_rel_allvisible, NULL);
Measure the number of all-visible pages for use in index-only scan costing. Add a column pg_class.relallvisible to remember the number of pages that were all-visible according to the visibility map as of the last VACUUM (or ANALYZE, or some other operations that update pg_class.relpages). Use relallvisible/relpages, instead of an arbitrary constant, to estimate how many heap page fetches can be avoided during an index-only scan. This is pretty primitive and will no doubt see refinements once we've acquired more field experience with the index-only scan mechanism, but it's way better than using a constant. Note: I had to adjust an underspecified query in the window.sql regression test, because it was changing answers when the plan changed to use an index-only scan. Some of the adjacent tests perhaps should be adjusted as well, but I didn't do that here.
2011-10-14 23:23:01 +02:00
if (new_rel_allvisible > new_rel_pages)
new_rel_allvisible = new_rel_pages;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/*
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
* Now actually update rel's pg_class entry.
*
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
* In principle new_live_tuples could be -1 indicating that we (still)
* don't know the tuple count. In practice that can't happen, since we
* scan every page that isn't skipped using the visibility map.
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
*/
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
vac_update_relstats(rel, new_rel_pages, vacrel->new_live_tuples,
new_rel_allvisible, vacrel->nindexes > 0,
vacrel->NewRelfrozenXid, vacrel->NewRelminMxid,
&frozenxid_updated, &minmulti_updated, false);
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
/*
pgstat: stats collector references in comments. Soon the stats collector will be no more, with statistics instead getting stored in shared memory. There are a lot of references to the stats collector in comments. This commit replaces most of these references with "cumulative statistics system", with the remaining ones getting replaced as part of subsequent commits. This is done separately from the - quite large - shared memory statistics patch to make review easier. Author: Andres Freund <andres@anarazel.de> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Reviewed-By: Thomas Munro <thomas.munro@gmail.com> Reviewed-By: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Discussion: https://postgr.es/m/20220303021600.hs34ghqcw6zcokdh@alap3.anarazel.de Discussion: https://postgr.es/m/20220308205351.2xcn6k4x5yivcxyd@alap3.anarazel.de
2022-04-06 22:56:06 +02:00
* Report results to the cumulative stats system, too.
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
*
pgstat: stats collector references in comments. Soon the stats collector will be no more, with statistics instead getting stored in shared memory. There are a lot of references to the stats collector in comments. This commit replaces most of these references with "cumulative statistics system", with the remaining ones getting replaced as part of subsequent commits. This is done separately from the - quite large - shared memory statistics patch to make review easier. Author: Andres Freund <andres@anarazel.de> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Reviewed-By: Thomas Munro <thomas.munro@gmail.com> Reviewed-By: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Discussion: https://postgr.es/m/20220303021600.hs34ghqcw6zcokdh@alap3.anarazel.de Discussion: https://postgr.es/m/20220308205351.2xcn6k4x5yivcxyd@alap3.anarazel.de
2022-04-06 22:56:06 +02:00
* Deliberately avoid telling the stats system about LP_DEAD items that
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
* remain in the table due to VACUUM bypassing index and heap vacuuming.
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* ANALYZE will consider the remaining LP_DEAD items to be dead "tuples".
* It seems like a good idea to err on the side of not vacuuming again too
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
* soon in cases where the failsafe prevented significant amounts of heap
* vacuuming.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
pgstat_report_vacuum(RelationGetRelid(rel),
rel->rd_rel->relisshared,
vacuumlazy.c: Clean up variable declarations. Move some of the heap_vacuum_rel() instrumentation related variables to the scope where they're actually needed. Also reorder some of the variable declarations at the start of heap_vacuum_rel() so that related variables appear together.
2022-04-02 19:33:21 +02:00
Max(vacrel->new_live_tuples, 0),
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
vacrel->recently_dead_tuples +
vacrel->missed_dead_tuples);
Add a generic command progress reporting facility. Using this facility, any utility command can report the target relation upon which it is operating, if there is one, and up to 10 64-bit counters; the intent of this is that users should be able to figure out what a utility command is doing without having to resort to ugly hacks like attaching strace to a backend. As a demonstration, this adds very crude reporting to lazy vacuum; we just report the target relation and nothing else. A forthcoming patch will make VACUUM report a bunch of additional data that will make this much more interesting. But this gets the basic framework in place. Vinayak Pokale, Rahila Syed, Amit Langote, Robert Haas, reviewed by Kyotaro Horiguchi, Jim Nasby, Thom Brown, Masahiko Sawada, Fujii Masao, and Masanori Oyama.
2016-03-09 18:08:58 +01:00
pgstat_progress_end_command();
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
if (instrument)
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
{
Improve logging of autovacuum I/O activity This adds some I/O stats to the logging of autovacuum (when the operation takes long enough that log_autovacuum_min_duration causes it to be logged), so that it is easier to tune. Notably, it adds buffer I/O counts (hits, misses, dirtied) and read and write rate. Authors: Greg Smith and Noah Misch
2011-11-25 16:10:46 +01:00
TimestampTz endtime = GetCurrentTimestamp();
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
if (verbose || params->log_min_duration == 0 ||
Improve logging of autovacuum I/O activity This adds some I/O stats to the logging of autovacuum (when the operation takes long enough that log_autovacuum_min_duration causes it to be logged), so that it is easier to tune. Notably, it adds buffer I/O counts (hits, misses, dirtied) and read and write rate. Authors: Greg Smith and Noah Misch
2011-11-25 16:10:46 +01:00
TimestampDifferenceExceeds(starttime, endtime,
Add log_min_autovacuum_duration per-table option This is useful to control autovacuum log volume, for situations where monitoring only a set of tables is necessary. Author: Michael Paquier Reviewed by: A team led by Naoya Anzai (also including Akira Kurosawa, Taiki Kondo, Huong Dangminh), Fujii Masao.
2015-04-03 16:55:50 +02:00
params->log_min_duration))
Improve logging of autovacuum I/O activity This adds some I/O stats to the logging of autovacuum (when the operation takes long enough that log_autovacuum_min_duration causes it to be logged), so that it is easier to tune. Notably, it adds buffer I/O counts (hits, misses, dirtied) and read and write rate. Authors: Greg Smith and Noah Misch
2011-11-25 16:10:46 +01:00
{
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
long secs_dur;
int usecs_dur;
vacuumlazy.c: Clean up variable declarations. Move some of the heap_vacuum_rel() instrumentation related variables to the scope where they're actually needed. Also reorder some of the variable declarations at the start of heap_vacuum_rel() so that related variables appear together.
2022-04-02 19:33:21 +02:00
WalUsage walusage;
Have VACUUM log number of skipped pages due to pins Author: Jim Nasby, some kibitzing by Heikki Linnankangas. Discussion leading to current behavior and precise wording fueled by thoughts from Robert Haas and Andres Freund.
2014-12-18 21:18:33 +01:00
StringInfoData buf;
In relevant log messages, indicate whether vacuums are aggressive. Kyotaro Horiguchi, reviewed Masahiko Sawada, David G. Johnston, Álvaro Herrera, and me. Grammar correction to the final posted patch by me. Discussion: http://postgr.es/m/20170329.124649.193656100.horiguchi.kyotaro@lab.ntt.co.jp
2017-10-26 12:35:34 +02:00
char *msgfmt;
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
int32 diff;
Fix multi-table VACUUM VERBOSE accounting. Per-backend global variables like VacuumPageHit are initialized once per VACUUM command. This was missed by commit 49c9d9fc, which unified VACUUM VERBOSE and autovacuum logging. As a result of that oversight, incorrect values were shown when multiple relations were processed by a single VACUUM VERBOSE command. Relations that happened to be processed later on would show "buffer usage:" values that incorrectly included buffer accesses made while processing earlier unrelated relations. The same accesses were counted multiple times. To fix, take initial values for the tracker variables at the start of heap_vacuum_rel(), and report delta values later on.
2022-04-16 00:48:39 +02:00
int64 PageHitOp = VacuumPageHit - StartPageHit,
PageMissOp = VacuumPageMiss - StartPageMiss,
PageDirtyOp = VacuumPageDirty - StartPageDirty;
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
double read_rate = 0,
write_rate = 0;
pgindent run for 9.5
2015-05-24 03:35:49 +02:00
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
TimestampDifference(starttime, endtime, &secs_dur, &usecs_dur);
Allow autovacuum to log WAL usage statistics. This commit allows autovacuum to log WAL usage statistics added by commit df3b181499. Author: Julien Rouhaud Reviewed-by: Dilip Kumar and Amit Kapila Discussion: https://postgr.es/m/CAB-hujrP8ZfUkvL5OYETipQwA=e3n7oqHFU=4ZLxWS_Cza3kQQ@mail.gmail.com
2020-04-06 12:54:51 +02:00
memset(&walusage, 0, sizeof(WalUsage));
Fix multi-table VACUUM VERBOSE accounting. Per-backend global variables like VacuumPageHit are initialized once per VACUUM command. This was missed by commit 49c9d9fc, which unified VACUUM VERBOSE and autovacuum logging. As a result of that oversight, incorrect values were shown when multiple relations were processed by a single VACUUM VERBOSE command. Relations that happened to be processed later on would show "buffer usage:" values that incorrectly included buffer accesses made while processing earlier unrelated relations. The same accesses were counted multiple times. To fix, take initial values for the tracker variables at the start of heap_vacuum_rel(), and report delta values later on.
2022-04-16 00:48:39 +02:00
WalUsageAccumDiff(&walusage, &pgWalUsage, &startwalusage);
Allow autovacuum to log WAL usage statistics. This commit allows autovacuum to log WAL usage statistics added by commit df3b181499. Author: Julien Rouhaud Reviewed-by: Dilip Kumar and Amit Kapila Discussion: https://postgr.es/m/CAB-hujrP8ZfUkvL5OYETipQwA=e3n7oqHFU=4ZLxWS_Cza3kQQ@mail.gmail.com
2020-04-06 12:54:51 +02:00
Have VACUUM log number of skipped pages due to pins Author: Jim Nasby, some kibitzing by Heikki Linnankangas. Discussion leading to current behavior and precise wording fueled by thoughts from Robert Haas and Andres Freund.
2014-12-18 21:18:33 +01:00
initStringInfo(&buf);
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
if (verbose)
{
/*
* Aggressiveness already reported earlier, in dedicated
* VACUUM VERBOSE ereport
*/
Assert(!params->is_wraparound);
msgfmt = _("finished vacuuming \"%s.%s.%s\": index scans: %d\n");
}
else if (params->is_wraparound)
Improve autovacuum logging for aggressive and anti-wraparound runs A log message was being generated when log_min_duration is reached for autovacuum on a given relation to indicate if it was an aggressive run, and missed the point of mentioning if it is doing an anti-wrapround run. The log message generated is improved so as one, both or no extra details are added depending on the option set. Author: Sergei Kornilov Reviewed-by: Masahiko Sawada, Michael Paquier Discussion: https://postgr.es/m/11587951532155118@sas1-19a94364928d.qloud-c.yandex.net
2018-09-14 00:35:39 +02:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/*
* While it's possible for a VACUUM to be both is_wraparound
* and !aggressive, that's just a corner-case -- is_wraparound
* implies aggressive. Produce distinct output for the corner
* case all the same, just in case.
*/
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
if (vacrel->aggressive)
Revert "Skip redundant anti-wraparound vacuums" This reverts commit 2aa6e33, that added a fast path to skip anti-wraparound and non-aggressive autovacuum jobs (these have no sense as anti-wraparound implies aggressive). With a cluster using a high amount of relations with a portion of them being heavily updated, this could cause autovacuum to lock down, with autovacuum workers attempting repeatedly those jobs on the same relations for the same database, that just kept being skipped. This lock down can be solved with a manual VACUUM FREEZE. Justin King has reported one environment where the issue happened, and Julien Rouhaud and I have been able to reproduce it in a second environment. With a very aggressive autovacuum_freeze_max_age, triggering those jobs with pgbench is a matter of minutes, and hitting the lock down is a lot harder (my local tests failed to do that). Note that anti-wraparound and non-aggressive jobs can only be triggered on a subset of shared catalogs: - pg_auth_members - pg_authid - pg_database - pg_replication_origin - pg_shseclabel - pg_subscription - pg_tablespace While the lock down was possible down to v12, the root cause of those jobs is a much older issue, which needs more analysis. Bonus thanks to Andres Freund for the discussion. Reported-by: Justin King Discussion: https://postgr.es/m/CAE39h22zPLrkH17GrkDgAYL3kbjvySYD1io+rtnAUFnaJJVS4g@mail.gmail.com Backpatch-through: 12
2020-03-31 01:27:47 +02:00
msgfmt = _("automatic aggressive vacuum to prevent wraparound of table \"%s.%s.%s\": index scans: %d\n");
else
msgfmt = _("automatic vacuum to prevent wraparound of table \"%s.%s.%s\": index scans: %d\n");
Improve autovacuum logging for aggressive and anti-wraparound runs A log message was being generated when log_min_duration is reached for autovacuum on a given relation to indicate if it was an aggressive run, and missed the point of mentioning if it is doing an anti-wrapround run. The log message generated is improved so as one, both or no extra details are added depending on the option set. Author: Sergei Kornilov Reviewed-by: Masahiko Sawada, Michael Paquier Discussion: https://postgr.es/m/11587951532155118@sas1-19a94364928d.qloud-c.yandex.net
2018-09-14 00:35:39 +02:00
}
In relevant log messages, indicate whether vacuums are aggressive. Kyotaro Horiguchi, reviewed Masahiko Sawada, David G. Johnston, Álvaro Herrera, and me. Grammar correction to the final posted patch by me. Discussion: http://postgr.es/m/20170329.124649.193656100.horiguchi.kyotaro@lab.ntt.co.jp
2017-10-26 12:35:34 +02:00
else
Improve autovacuum logging for aggressive and anti-wraparound runs A log message was being generated when log_min_duration is reached for autovacuum on a given relation to indicate if it was an aggressive run, and missed the point of mentioning if it is doing an anti-wrapround run. The log message generated is improved so as one, both or no extra details are added depending on the option set. Author: Sergei Kornilov Reviewed-by: Masahiko Sawada, Michael Paquier Discussion: https://postgr.es/m/11587951532155118@sas1-19a94364928d.qloud-c.yandex.net
2018-09-14 00:35:39 +02:00
{
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
if (vacrel->aggressive)
Improve autovacuum logging for aggressive and anti-wraparound runs A log message was being generated when log_min_duration is reached for autovacuum on a given relation to indicate if it was an aggressive run, and missed the point of mentioning if it is doing an anti-wrapround run. The log message generated is improved so as one, both or no extra details are added depending on the option set. Author: Sergei Kornilov Reviewed-by: Masahiko Sawada, Michael Paquier Discussion: https://postgr.es/m/11587951532155118@sas1-19a94364928d.qloud-c.yandex.net
2018-09-14 00:35:39 +02:00
msgfmt = _("automatic aggressive vacuum of table \"%s.%s.%s\": index scans: %d\n");
else
msgfmt = _("automatic vacuum of table \"%s.%s.%s\": index scans: %d\n");
}
In relevant log messages, indicate whether vacuums are aggressive. Kyotaro Horiguchi, reviewed Masahiko Sawada, David G. Johnston, Álvaro Herrera, and me. Grammar correction to the final posted patch by me. Discussion: http://postgr.es/m/20170329.124649.193656100.horiguchi.kyotaro@lab.ntt.co.jp
2017-10-26 12:35:34 +02:00
appendStringInfo(&buf, msgfmt,
vacuumlazy.c: Save get_database_name() in vacrel. This brings dbname strings in line with namespace and relation name strings. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkQ1TKU-DdNvnGeL870di3+CU1UTo-7nw7xFDpVE-XGjA@mail.gmail.com
2023-01-03 20:48:47 +01:00
vacrel->dbname,
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->relnamespace,
vacrel->relname,
vacrel->num_index_scans);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
appendStringInfo(&buf, _("pages: %u removed, %u remain, %u scanned (%.2f%% of total)\n"),
vacuumlazy.c: Rename state field for consistency. Rename pages_removed to removed_pages, for consistency with nearby vacrel fields.
2022-01-29 02:41:09 +01:00
vacrel->removed_pages,
vacuumlazy.c: Clean up variable declarations. Move some of the heap_vacuum_rel() instrumentation related variables to the scope where they're actually needed. Also reorder some of the variable declarations at the start of heap_vacuum_rel() so that related variables appear together.
2022-04-02 19:33:21 +02:00
new_rel_pages,
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
vacrel->scanned_pages,
VACUUM VERBOSE: tweak scanned_pages logic. Commit 872770fd6c taught VACUUM VERBOSE and autovacuum logging to display the total number of pages scanned by VACUUM. This information was also displayed as a percentage of rel_pages in parenthesis, which makes it easy to spot trends over time and across tables. The instrumentation displayed "0 scanned (0.00% of total)" for totally empty tables. Tweak the instrumentation: have it show "0 scanned (100.00% of total)" for empty tables instead. This approach is clearer and more consistent.
2022-03-13 21:07:49 +01:00
orig_rel_pages == 0 ? 100.0 :
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
100.0 * vacrel->scanned_pages / orig_rel_pages);
Have VACUUM log number of skipped pages due to pins Author: Jim Nasby, some kibitzing by Heikki Linnankangas. Discussion leading to current behavior and precise wording fueled by thoughts from Robert Haas and Andres Freund.
2014-12-18 21:18:33 +01:00
appendStringInfo(&buf,
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
_("tuples: %lld removed, %lld remain, %lld are dead but not yet removable\n"),
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
(long long) vacrel->tuples_deleted,
(long long) vacrel->new_rel_tuples,
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
(long long) vacrel->recently_dead_tuples);
if (vacrel->missed_dead_tuples > 0)
appendStringInfo(&buf,
_("tuples missed: %lld dead from %u pages not removed due to cleanup lock contention\n"),
(long long) vacrel->missed_dead_tuples,
vacrel->missed_dead_pages);
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
diff = (int32) (ReadNextTransactionId() -
vacrel->cutoffs.OldestXmin);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
appendStringInfo(&buf,
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
_("removable cutoff: %u, which was %d XIDs old when operation ended\n"),
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
vacrel->cutoffs.OldestXmin, diff);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
if (frozenxid_updated)
{
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
diff = (int32) (vacrel->NewRelfrozenXid -
vacrel->cutoffs.relfrozenxid);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
appendStringInfo(&buf,
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
_("new relfrozenxid: %u, which is %d XIDs ahead of previous value\n"),
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
vacrel->NewRelfrozenXid, diff);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
}
if (minmulti_updated)
{
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
diff = (int32) (vacrel->NewRelminMxid -
vacrel->cutoffs.relminmxid);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
appendStringInfo(&buf,
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
_("new relminmxid: %u, which is %d MXIDs ahead of previous value\n"),
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
vacrel->NewRelminMxid, diff);
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
}
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
appendStringInfo(&buf, _("frozen: %u pages from table (%.2f%% of total) had %lld tuples frozen\n"),
vacrel->frozen_pages,
orig_rel_pages == 0 ? 100.0 :
100.0 * vacrel->frozen_pages / orig_rel_pages,
(long long) vacrel->tuples_frozen);
VACUUM VERBOSE: Show dead items for an empty table. Be consistent about the lines that VACUUM VERBOSE outputs by including an "index scan not needed: " line for completely empty tables. This makes the output more readable, especially with multiple distinct VACUUM operations processed by the same VACUUM command. It's also more consistent; even empty tables can use the failsafe, which wasn't reported in the standard way until now. Follow-up to commit 6e20f460, which taught VACUUM VERBOSE to be more consistent about reporting on scanned pages with empty tables.
2022-04-15 23:20:56 +02:00
if (vacrel->do_index_vacuuming)
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
{
VACUUM VERBOSE: Show dead items for an empty table. Be consistent about the lines that VACUUM VERBOSE outputs by including an "index scan not needed: " line for completely empty tables. This makes the output more readable, especially with multiple distinct VACUUM operations processed by the same VACUUM command. It's also more consistent; even empty tables can use the failsafe, which wasn't reported in the standard way until now. Follow-up to commit 6e20f460, which taught VACUUM VERBOSE to be more consistent about reporting on scanned pages with empty tables.
2022-04-15 23:20:56 +02:00
if (vacrel->nindexes == 0 || vacrel->num_index_scans == 0)
appendStringInfoString(&buf, _("index scan not needed: "));
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
else
VACUUM VERBOSE: Show dead items for an empty table. Be consistent about the lines that VACUUM VERBOSE outputs by including an "index scan not needed: " line for completely empty tables. This makes the output more readable, especially with multiple distinct VACUUM operations processed by the same VACUUM command. It's also more consistent; even empty tables can use the failsafe, which wasn't reported in the standard way until now. Follow-up to commit 6e20f460, which taught VACUUM VERBOSE to be more consistent about reporting on scanned pages with empty tables.
2022-04-15 23:20:56 +02:00
appendStringInfoString(&buf, _("index scan needed: "));
msgfmt = _("%u pages from table (%.2f%% of total) had %lld dead item identifiers removed\n");
}
else
{
Make vacuum failsafe_active globally visible While vacuuming a table in failsafe mode, VacuumCostActive should not be re-enabled. This currently isn't a problem because vacuum cost parameters are only refreshed in between vacuuming tables and failsafe status is reset for every table. In preparation for allowing vacuum cost parameters to be updated more frequently, elevate LVRelState->failsafe_active to a global, VacuumFailsafeActive, which will be checked when determining whether or not to re-enable vacuum cost-related delays. Author: Melanie Plageman <melanieplageman@gmail.com> Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Daniel Gustafsson <daniel@yesql.se> Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Reviewed-by: Robert Haas <robertmhaas@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/CAAKRu_ZngzqnEODc7LmS1NH04Kt6Y9huSjz5pp7%2BDXhrjDA0gw%40mail.gmail.com
2023-04-07 00:54:08 +02:00
if (!VacuumFailsafeActive)
VACUUM VERBOSE: Show dead items for an empty table. Be consistent about the lines that VACUUM VERBOSE outputs by including an "index scan not needed: " line for completely empty tables. This makes the output more readable, especially with multiple distinct VACUUM operations processed by the same VACUUM command. It's also more consistent; even empty tables can use the failsafe, which wasn't reported in the standard way until now. Follow-up to commit 6e20f460, which taught VACUUM VERBOSE to be more consistent about reporting on scanned pages with empty tables.
2022-04-15 23:20:56 +02:00
appendStringInfoString(&buf, _("index scan bypassed: "));
else
appendStringInfoString(&buf, _("index scan bypassed by failsafe: "));
msgfmt = _("%u pages from table (%.2f%% of total) have %lld dead item identifiers\n");
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
}
VACUUM VERBOSE: Show dead items for an empty table. Be consistent about the lines that VACUUM VERBOSE outputs by including an "index scan not needed: " line for completely empty tables. This makes the output more readable, especially with multiple distinct VACUUM operations processed by the same VACUUM command. It's also more consistent; even empty tables can use the failsafe, which wasn't reported in the standard way until now. Follow-up to commit 6e20f460, which taught VACUUM VERBOSE to be more consistent about reporting on scanned pages with empty tables.
2022-04-15 23:20:56 +02:00
appendStringInfo(&buf, msgfmt,
vacrel->lpdead_item_pages,
orig_rel_pages == 0 ? 100.0 :
100.0 * vacrel->lpdead_item_pages / orig_rel_pages,
(long long) vacrel->lpdead_items);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
for (int i = 0; i < vacrel->nindexes; i++)
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
IndexBulkDeleteResult *istat = vacrel->indstats[i];
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (!istat)
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
continue;
appendStringInfo(&buf,
Reword slightly logs generated for index stats in autovacuum Using "remain" is confusing, as it implies that the index file can shrink. Instead, use "in total". Per discussion with Peter Geoghegan. Discussion: https://postgr.es/m/CAH2-WzkYgHZzpGOwR14CScJsjaQpvJrEkEfkh_=wGhzLb=yVdQ@mail.gmail.com
2021-03-24 01:36:03 +01:00
_("index \"%s\": pages: %u in total, %u newly deleted, %u currently deleted, %u reusable\n"),
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
indnames[i],
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
istat->num_pages,
istat->pages_newly_deleted,
istat->pages_deleted,
istat->pages_free);
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
}
Improve logging of auto-vacuum and auto-analyze When logging auto-vacuum and auto-analyze activity, include the I/O timing if track_io_timing is enabled. Also, for auto-analyze, add the read rate and the dirty rate, similar to how that information has historically been logged for auto-vacuum. Stephen Frost and Jakub Wartak Reviewed-By: Heikki Linnakangas, Tomas Vondra Discussion: https://www.postgresql.org/message-id/VI1PR0701MB69603A433348EDCF783C6ECBF6EF0%40VI1PR0701MB6960.eurprd07.prod.outlook.com
2021-03-16 19:46:48 +01:00
if (track_io_timing)
{
track_io_timing logging: Don't special case 0 ms. Adjust track_io_timing related logging code added by commit 94d13d474d. Make it consistent with other nearby autovacuum and autoanalyze logging code by removing logic that suppressed zero millisecond outputs. log_autovacuum_min_duration log output now reliably shows "read:" and "write:" millisecond-based values in its report (when track_io_timing is enabled). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Stephen Frost <sfrost@snowman.net> Discussion: https://postgr.es/m/CAH2-WznW0FNxSVQMSRazAMYNfZ6DR_gr5WE78hc6E1CBkkJpzw@mail.gmail.com Backpatch: 14-, where the track_io_timing logging was introduced.
2021-08-27 22:34:00 +02:00
double read_ms = (double) (pgStatBlockReadTime - startreadtime) / 1000;
double write_ms = (double) (pgStatBlockWriteTime - startwritetime) / 1000;
appendStringInfo(&buf, _("I/O timings: read: %.3f ms, write: %.3f ms\n"),
read_ms, write_ms);
Improve logging of auto-vacuum and auto-analyze When logging auto-vacuum and auto-analyze activity, include the I/O timing if track_io_timing is enabled. Also, for auto-analyze, add the read rate and the dirty rate, similar to how that information has historically been logged for auto-vacuum. Stephen Frost and Jakub Wartak Reviewed-By: Heikki Linnakangas, Tomas Vondra Discussion: https://www.postgresql.org/message-id/VI1PR0701MB69603A433348EDCF783C6ECBF6EF0%40VI1PR0701MB6960.eurprd07.prod.outlook.com
2021-03-16 19:46:48 +01:00
}
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
if (secs_dur > 0 || usecs_dur > 0)
{
Fix multi-table VACUUM VERBOSE accounting. Per-backend global variables like VacuumPageHit are initialized once per VACUUM command. This was missed by commit 49c9d9fc, which unified VACUUM VERBOSE and autovacuum logging. As a result of that oversight, incorrect values were shown when multiple relations were processed by a single VACUUM VERBOSE command. Relations that happened to be processed later on would show "buffer usage:" values that incorrectly included buffer accesses made while processing earlier unrelated relations. The same accesses were counted multiple times. To fix, take initial values for the tracker variables at the start of heap_vacuum_rel(), and report delta values later on.
2022-04-16 00:48:39 +02:00
read_rate = (double) BLCKSZ * PageMissOp / (1024 * 1024) /
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
(secs_dur + usecs_dur / 1000000.0);
Fix multi-table VACUUM VERBOSE accounting. Per-backend global variables like VacuumPageHit are initialized once per VACUUM command. This was missed by commit 49c9d9fc, which unified VACUUM VERBOSE and autovacuum logging. As a result of that oversight, incorrect values were shown when multiple relations were processed by a single VACUUM VERBOSE command. Relations that happened to be processed later on would show "buffer usage:" values that incorrectly included buffer accesses made while processing earlier unrelated relations. The same accesses were counted multiple times. To fix, take initial values for the tracker variables at the start of heap_vacuum_rel(), and report delta values later on.
2022-04-16 00:48:39 +02:00
write_rate = (double) BLCKSZ * PageDirtyOp / (1024 * 1024) /
Adjust VACUUM's removable cutoff log message. The age of OldestXmin (a.k.a. "removable cutoff") when VACUUM ends often indicates the approximate number of XIDs consumed while VACUUM ran. However, there is at least one important exception: the cutoff could be held back by a snapshot that was acquired before our VACUUM even began. Successive VACUUM operations may even use exactly the same old cutoff in extreme cases involving long held snapshots. The log messages that described how removable cutoff aged (which were added by commit 872770fd) created the impression that we were reporting on how VACUUM's usable cutoff advanced while VACUUM ran, which was misleading in these extreme cases. Fix by using a more general wording. Per gripe from Tom Lane. In passing, relocate related instrumentation code for clarity. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/1643035.1650035653@sss.pgh.pa.us
2022-04-15 22:21:43 +02:00
(secs_dur + usecs_dur / 1000000.0);
}
Reorder log_autovacuum_min_duration log output. This order seems more natural. It starts with details that are particular to heap and index data structures, and ends with system-level costs incurred during the autovacuum worker's VACUUM/ANALYZE operation. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkzxK6ahA9xxsOftRtBX_R0swuHZsvo4QUbak1Bz7hb7Q@mail.gmail.com Backpatch: 14-, which enhanced the log output in various ways.
2021-08-27 22:08:41 +02:00
appendStringInfo(&buf, _("avg read rate: %.3f MB/s, avg write rate: %.3f MB/s\n"),
read_rate, write_rate);
appendStringInfo(&buf,
_("buffer usage: %lld hits, %lld misses, %lld dirtied\n"),
Fix multi-table VACUUM VERBOSE accounting. Per-backend global variables like VacuumPageHit are initialized once per VACUUM command. This was missed by commit 49c9d9fc, which unified VACUUM VERBOSE and autovacuum logging. As a result of that oversight, incorrect values were shown when multiple relations were processed by a single VACUUM VERBOSE command. Relations that happened to be processed later on would show "buffer usage:" values that incorrectly included buffer accesses made while processing earlier unrelated relations. The same accesses were counted multiple times. To fix, take initial values for the tracker variables at the start of heap_vacuum_rel(), and report delta values later on.
2022-04-16 00:48:39 +02:00
(long long) PageHitOp,
(long long) PageMissOp,
(long long) PageDirtyOp);
Allow autovacuum to log WAL usage statistics. This commit allows autovacuum to log WAL usage statistics added by commit df3b181499. Author: Julien Rouhaud Reviewed-by: Dilip Kumar and Amit Kapila Discussion: https://postgr.es/m/CAB-hujrP8ZfUkvL5OYETipQwA=e3n7oqHFU=4ZLxWS_Cza3kQQ@mail.gmail.com
2020-04-06 12:54:51 +02:00
appendStringInfo(&buf,
Reorder log_autovacuum_min_duration log output. This order seems more natural. It starts with details that are particular to heap and index data structures, and ends with system-level costs incurred during the autovacuum worker's VACUUM/ANALYZE operation. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkzxK6ahA9xxsOftRtBX_R0swuHZsvo4QUbak1Bz7hb7Q@mail.gmail.com Backpatch: 14-, which enhanced the log output in various ways.
2021-08-27 22:08:41 +02:00
_("WAL usage: %lld records, %lld full page images, %llu bytes\n"),
Change data type of counters in BufferUsage and WalUsage from long to int64. Previously long was used as the data type for some counters in BufferUsage and WalUsage. But long is only four byte, e.g., on Windows, and it's entirely possible to wrap a four byte counter. For example, emitting more than four billion WAL records in one transaction isn't actually particularly rare. To avoid the overflows of those counters, this commit changes the data type of them from long to int64. Suggested-by: Andres Freund Author: Masahiro Ikeda Reviewed-by: Fujii Masao Discussion: https://postgr.es/m/20201221211650.k7b53tcnadrciqjo@alap3.anarazel.de Discussion: https://postgr.es/m/af0964ac-7080-1984-dc23-513754987716@oss.nttdata.com
2021-05-12 02:56:34 +02:00
(long long) walusage.wal_records,
(long long) walusage.wal_fpi,
Message fixes and style improvements
2020-09-14 06:42:07 +02:00
(unsigned long long) walusage.wal_bytes);
Reorder log_autovacuum_min_duration log output. This order seems more natural. It starts with details that are particular to heap and index data structures, and ends with system-level costs incurred during the autovacuum worker's VACUUM/ANALYZE operation. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkzxK6ahA9xxsOftRtBX_R0swuHZsvo4QUbak1Bz7hb7Q@mail.gmail.com Backpatch: 14-, which enhanced the log output in various ways.
2021-08-27 22:08:41 +02:00
appendStringInfo(&buf, _("system usage: %s"), pg_rusage_show(&ru0));
Have VACUUM log number of skipped pages due to pins Author: Jim Nasby, some kibitzing by Heikki Linnankangas. Discussion leading to current behavior and precise wording fueled by thoughts from Robert Haas and Andres Freund.
2014-12-18 21:18:33 +01:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ereport(verbose ? INFO : LOG,
Have VACUUM log number of skipped pages due to pins Author: Jim Nasby, some kibitzing by Heikki Linnankangas. Discussion leading to current behavior and precise wording fueled by thoughts from Robert Haas and Andres Freund.
2014-12-18 21:18:33 +01:00
(errmsg_internal("%s", buf.data)));
pfree(buf.data);
Improve logging of autovacuum I/O activity This adds some I/O stats to the logging of autovacuum (when the operation takes long enough that log_autovacuum_min_duration causes it to be logged), so that it is easier to tune. Notably, it adds buffer I/O counts (hits, misses, dirtied) and read and write rate. Authors: Greg Smith and Noah Misch
2011-11-25 16:10:46 +01:00
}
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
}
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
/* Cleanup index statistics and index names */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
for (int i = 0; i < vacrel->nindexes; i++)
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (vacrel->indstats[i])
pfree(vacrel->indstats[i]);
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
if (instrument)
Add per-index stats information in verbose logs of autovacuum Once a relation's autovacuum is completed, the logs include more information about this relation state if the threshold of log_autovacuum_min_duration (or its relation option) is reached, with for example contents about the statistics of the VACUUM operation for the relation, WAL and system usage. This commit adds more information about the statistics of the relation's indexes, with one line of logs generated for each index. The index stats were already calculated, but not printed in the context of autovacuum yet. While on it, some refactoring is done to keep track of the index statistics directly within LVRelStats, simplifying some routines related to parallel VACUUMs. Author: Masahiko Sawada Reviewed-by: Michael Paquier, Euler Taveira Discussion: https://postgr.es/m/CAD21AoAy6SxHiTivh5yAPJSUE4S=QRPpSZUdafOSz0R+fRcM6Q@mail.gmail.com
2021-03-23 05:25:14 +01:00
pfree(indnames[i]);
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
/*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* lazy_scan_heap() -- workhorse function for VACUUM
*
* This routine prunes each page in the heap, and considers the need to
* freeze remaining tuples with storage (not including pages that can be
* skipped using the visibility map). Also performs related maintenance
* of the FSM and visibility map. These steps all take place during an
* initial pass over the target heap relation.
*
* Also invokes lazy_vacuum_all_indexes to vacuum indexes, which largely
* consists of deleting index tuples that point to LP_DEAD items left in
* heap pages following pruning. Earlier initial pass over the heap will
* have collected the TIDs whose index tuples need to be removed.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* Finally, invokes lazy_vacuum_heap_rel to vacuum heap pages, which
* largely consists of marking LP_DEAD items (from collected TID array)
* as LP_UNUSED. This has to happen in a second, final pass over the
* heap, to preserve a basic invariant that all index AMs rely on: no
* extant index tuple can ever be allowed to contain a TID that points to
* an LP_UNUSED line pointer in the heap. We must disallow premature
* recycling of line pointers to avoid index scans that get confused
* about which TID points to which tuple immediately after recycling.
* (Actually, this isn't a concern when target heap relation happens to
* have no indexes, which allows us to safely apply the one-pass strategy
* as an optimization).
Improve vacuum code to track minimum Xids per table instead of per database. To this end, add a couple of columns to pg_class, relminxid and relvacuumxid, based on which we calculate the pg_database columns after each vacuum. We now force all databases to be vacuumed, even template ones. A backend noticing too old a database (meaning pg_database.datminxid is in danger of falling behind Xid wraparound) will signal the postmaster, which in turn will start an autovacuum iteration to process the offending database. In principle this is only there to cope with frozen (non-connectable) databases without forcing users to set them to connectable, but it could force regular user database to go through a database-wide vacuum at any time. Maybe we should warn users about this somehow. Of course the real solution will be to use autovacuum all the time ;-) There are some additional improvements we could have in this area: for example the vacuum code could be smarter about not updating pg_database for each table when called by autovacuum, and do it only once the whole autovacuum iteration is done. I updated the system catalogs documentation, but I didn't modify the maintenance section. Also having some regression tests for this would be nice but it's not really a very straightforward thing to do. Catalog version bumped due to system catalog changes.
2006-07-10 18:20:52 +02:00
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* In practice we often have enough space to fit all TIDs, and so won't
* need to call lazy_vacuum more than once, after our initial pass over
* the heap has totally finished. Otherwise things are slightly more
* complicated: our "initial pass" over the heap applies only to those
* pages that were pruned before we needed to call lazy_vacuum, and our
* "final pass" over the heap only vacuums these same heap pages.
* However, we process indexes in full every time lazy_vacuum is called,
* which makes index processing very inefficient when memory is in short
* supply.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
static void
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
lazy_scan_heap(LVRelState *vacrel)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
vacuumlazy.c: Standardize rel_pages terminology. VACUUM's rel_pages field indicates the size of the target heap rel just after the table_relation_vacuum() operation began. There are specific expectations around how rel_pages can be related to other nearby state. In particular, the range of rel_pages must contain every tuple in the relation whose tuple headers might contain an XID < OldestXmin. Consistently refer to the field as rel_pages to make this clearer and more discoverable. This is follow-up work to commit 73f6ec3d from earlier today. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20220311031351.sbge5m2bpvy2ttxg@alap3.anarazel.de
2022-03-12 22:20:45 +01:00
BlockNumber rel_pages = vacrel->rel_pages,
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
blkno,
next_unskippable_block,
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
next_fsm_block_to_vacuum = 0;
VacDeadItems *dead_items = vacrel->dead_items;
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
Buffer vmbuffer = InvalidBuffer;
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
bool next_unskippable_allvis,
skipping_current_range;
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
const int initprog_index[] = {
PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_TOTAL_HEAP_BLKS,
PROGRESS_VACUUM_MAX_DEAD_TUPLES
};
int64 initprog_val[3];
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
/* Report that we're scanning the heap, advertising total # of blocks */
initprog_val[0] = PROGRESS_VACUUM_PHASE_SCAN_HEAP;
vacuumlazy.c: Standardize rel_pages terminology. VACUUM's rel_pages field indicates the size of the target heap rel just after the table_relation_vacuum() operation began. There are specific expectations around how rel_pages can be related to other nearby state. In particular, the range of rel_pages must contain every tuple in the relation whose tuple headers might contain an XID < OldestXmin. Consistently refer to the field as rel_pages to make this clearer and more discoverable. This is follow-up work to commit 73f6ec3d from earlier today. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20220311031351.sbge5m2bpvy2ttxg@alap3.anarazel.de
2022-03-12 22:20:45 +01:00
initprog_val[1] = rel_pages;
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
initprog_val[2] = dead_items->max_items;
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
pgstat_progress_update_multi_param(3, initprog_index, initprog_val);
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
/* Set up an initial range of skippable blocks using the visibility map */
next_unskippable_block = lazy_scan_skip(vacrel, &vmbuffer, 0,
&next_unskippable_allvis,
&skipping_current_range);
vacuumlazy.c: Standardize rel_pages terminology. VACUUM's rel_pages field indicates the size of the target heap rel just after the table_relation_vacuum() operation began. There are specific expectations around how rel_pages can be related to other nearby state. In particular, the range of rel_pages must contain every tuple in the relation whose tuple headers might contain an XID < OldestXmin. Consistently refer to the field as rel_pages to make this clearer and more discoverable. This is follow-up work to commit 73f6ec3d from earlier today. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20220311031351.sbge5m2bpvy2ttxg@alap3.anarazel.de
2022-03-12 22:20:45 +01:00
for (blkno = 0; blkno < rel_pages; blkno++)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Buffer buf;
Page page;
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
bool all_visible_according_to_vm;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
LVPagePruneState prunestate;
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
Don't vacuum all-frozen pages. Commit a892234f830e832110f63fc0a2afce2fb21d1584 gave us enough infrastructure to avoid vacuuming pages where every tuple on the page is already frozen. So, replace the notion of a scan_all or whole-table vacuum with the less onerous notion of an "aggressive" vacuum, which will pages that are all-visible, but still skip those that are all-frozen. This should greatly reduce the cost of anti-wraparound vacuuming on large clusters where the majority of data is never touched between one cycle and the next, because we'll no longer have to read all of those pages only to find out that we don't need to do anything with them. Patch by me, reviewed by Masahiko Sawada.
2016-03-10 22:12:10 +01:00
if (blkno == next_unskippable_block)
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
{
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
/*
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
* Can't skip this page safely. Must scan the page. But
* determine the next skippable range after the page first.
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
*/
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
all_visible_according_to_vm = next_unskippable_allvis;
next_unskippable_block = lazy_scan_skip(vacrel, &vmbuffer,
blkno + 1,
&next_unskippable_allvis,
&skipping_current_range);
Don't vacuum all-frozen pages. Commit a892234f830e832110f63fc0a2afce2fb21d1584 gave us enough infrastructure to avoid vacuuming pages where every tuple on the page is already frozen. So, replace the notion of a scan_all or whole-table vacuum with the less onerous notion of an "aggressive" vacuum, which will pages that are all-visible, but still skip those that are all-frozen. This should greatly reduce the cost of anti-wraparound vacuuming on large clusters where the majority of data is never touched between one cycle and the next, because we'll no longer have to read all of those pages only to find out that we don't need to do anything with them. Patch by me, reviewed by Masahiko Sawada.
2016-03-10 22:12:10 +01:00
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
Assert(next_unskippable_block >= blkno + 1);
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
}
else
{
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
/* Last page always scanned (may need to set nonempty_pages) */
Assert(blkno < rel_pages - 1);
if (skipping_current_range)
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
continue;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
/* Current range is too small to skip -- just scan the page */
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
all_visible_according_to_vm = true;
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
vacrel->scanned_pages++;
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
/* Report as block scanned, update error traceback information */
pgstat_progress_update_param(PROGRESS_VACUUM_HEAP_BLKS_SCANNED, blkno);
update_vacuum_error_info(vacrel, NULL, VACUUM_ERRCB_PHASE_SCAN_HEAP,
blkno, InvalidOffsetNumber);
vacuum_delay_point();
Consider triggering VACUUM failsafe during scan. The wraparound failsafe mechanism added by commit 1e55e7d1 handled the one-pass strategy case (i.e. the "table has no indexes" case) by adding a dedicated failsafe check. This made up for the fact that the usual one-pass checks inside lazy_vacuum_all_indexes() cannot ever be reached during a one-pass strategy VACUUM. This approach failed to account for two-pass VACUUMs that opt out of index vacuuming up-front. The INDEX_CLEANUP off case in the only case that works like that. Fix this by performing a failsafe check every 4GB during the first scan of the heap, regardless of the details of the VACUUM. This eliminates the special case, and will make the failsafe trigger more reliably. Author: Peter Geoghegan <pg@bowt.ie> Reported-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/20210424002921.pb3t7h6frupdqnkp@alap3.anarazel.de
2021-05-25 02:14:02 +02:00
/*
* Regularly check if wraparound failsafe should trigger.
*
* There is a similar check inside lazy_vacuum_all_indexes(), but
* relfrozenxid might start to look dangerously old before we reach
* that point. This check also provides failsafe coverage for the
Remove overzealous VACUUM failsafe assertions. The failsafe can trigger when index processing is already disabled. This can happen when VACUUM's INDEX_CLEANUP parameter is "off" and the failsafe happens to trigger. Remove assertions that assume that index processing is directly tied to the failsafe. Oversight in commit c242baa4, which made it possible for the failsafe to trigger in a two-pass strategy VACUUM that has yet to make its first call to lazy_vacuum_all_indexes().
2021-06-21 03:14:00 +02:00
* one-pass strategy, and the two-pass strategy with the index_cleanup
* param set to 'off'.
Consider triggering VACUUM failsafe during scan. The wraparound failsafe mechanism added by commit 1e55e7d1 handled the one-pass strategy case (i.e. the "table has no indexes" case) by adding a dedicated failsafe check. This made up for the fact that the usual one-pass checks inside lazy_vacuum_all_indexes() cannot ever be reached during a one-pass strategy VACUUM. This approach failed to account for two-pass VACUUMs that opt out of index vacuuming up-front. The INDEX_CLEANUP off case in the only case that works like that. Fix this by performing a failsafe check every 4GB during the first scan of the heap, regardless of the details of the VACUUM. This eliminates the special case, and will make the failsafe trigger more reliably. Author: Peter Geoghegan <pg@bowt.ie> Reported-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/20210424002921.pb3t7h6frupdqnkp@alap3.anarazel.de
2021-05-25 02:14:02 +02:00
*/
Use scanned_pages to decide when to failsafe check. Perform a failsafe check every time VACUUM's first heap scan scans a further FAILSAFE_EVERY_PAGES pages, rather than using an approach based on the number of physical blocks that our current blkno is from the blkno at the time of the previous failsafe check. That way VACUUM will perform a failsafe check every time it has scanned a uniform number of pages, without it mattering when or how VACUUM skipped pages using the visibility map. Sami Imseih, with changes to FAILSAFE_EVERY_PAGES comments added by me. Author: Sami Imseih <simseih@amazon.com> Reviewed-By: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/401CE010-4049-4B94-9961-0B610A5D254D%40amazon.com
2022-12-22 19:41:40 +01:00
if (vacrel->scanned_pages % FAILSAFE_EVERY_PAGES == 0)
Consider triggering VACUUM failsafe during scan. The wraparound failsafe mechanism added by commit 1e55e7d1 handled the one-pass strategy case (i.e. the "table has no indexes" case) by adding a dedicated failsafe check. This made up for the fact that the usual one-pass checks inside lazy_vacuum_all_indexes() cannot ever be reached during a one-pass strategy VACUUM. This approach failed to account for two-pass VACUUMs that opt out of index vacuuming up-front. The INDEX_CLEANUP off case in the only case that works like that. Fix this by performing a failsafe check every 4GB during the first scan of the heap, regardless of the details of the VACUUM. This eliminates the special case, and will make the failsafe trigger more reliably. Author: Peter Geoghegan <pg@bowt.ie> Reported-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/20210424002921.pb3t7h6frupdqnkp@alap3.anarazel.de
2021-05-25 02:14:02 +02:00
lazy_check_wraparound_failsafe(vacrel);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Consider if we definitely have enough space to process TIDs on page
* already. If we are close to overrunning the available space for
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* dead_items TIDs, pause and do a cycle of vacuuming before we tackle
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* this page.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
Assert(dead_items->max_items >= MaxHeapTuplesPerPage);
if (dead_items->max_items - dead_items->num_items < MaxHeapTuplesPerPage)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
/*
* Before beginning index vacuuming, we release any pin we may
* hold on the visibility map page. This isn't necessary for
* correctness, but we do it anyway to avoid holding the pin
* across a lengthy, unrelated operation.
*/
if (BufferIsValid(vmbuffer))
{
ReleaseBuffer(vmbuffer);
vmbuffer = InvalidBuffer;
}
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
/* Perform a round of index and heap vacuuming */
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
vacrel->consider_bypass_optimization = false;
lazy_vacuum(vacrel);
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
/*
* Vacuum the Free Space Map to make newly-freed space visible on
* upper-level FSM pages. Note we have not yet processed blkno.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
FreeSpaceMapVacuumRange(vacrel->rel, next_fsm_block_to_vacuum,
blkno);
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
next_fsm_block_to_vacuum = blkno;
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
/* Report that we are once again scanning the heap */
pgstat_progress_update_param(PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_PHASE_SCAN_HEAP);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
/*
* Pin the visibility map page in case we need to mark the page
* all-visible. In most cases this will be very cheap, because we'll
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
* already have the correct page pinned anyway.
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
visibilitymap_pin(vacrel->rel, blkno, &vmbuffer);
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
* We need a buffer cleanup lock to prune HOT chains and defragment
* the page in lazy_scan_prune. But when it's not possible to acquire
* a cleanup lock right away, we may be able to settle for reduced
* processing using lazy_scan_noprune.
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
buf = ReadBufferExtended(vacrel->rel, MAIN_FORKNUM, blkno, RBM_NORMAL,
vacrel->bstrategy);
page = BufferGetPage(buf);
Make VACUUM avoid waiting for a cleanup lock, where possible. In a regular VACUUM, it's OK to skip pages for which a cleanup lock isn't immediately available; the next VACUUM will deal with them. If we're scanning the entire relation to advance relfrozenxid, we might need to wait, but only if there are tuples on the page that actually require freezing. These changes should greatly reduce the incidence of of vacuum processes getting "stuck". Simon Riggs and Robert Haas
2011-11-08 03:39:40 +01:00
if (!ConditionalLockBufferForCleanup(buf))
{
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
bool hastup,
recordfreespace;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
Make VACUUM avoid waiting for a cleanup lock, where possible. In a regular VACUUM, it's OK to skip pages for which a cleanup lock isn't immediately available; the next VACUUM will deal with them. If we're scanning the entire relation to advance relfrozenxid, we might need to wait, but only if there are tuples on the page that actually require freezing. These changes should greatly reduce the incidence of of vacuum processes getting "stuck". Simon Riggs and Robert Haas
2011-11-08 03:39:40 +01:00
LockBuffer(buf, BUFFER_LOCK_SHARE);
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Check for new or empty pages before lazy_scan_noprune call */
if (lazy_scan_new_or_empty(vacrel, buf, blkno, page, true,
vmbuffer))
Make VACUUM avoid waiting for a cleanup lock, where possible. In a regular VACUUM, it's OK to skip pages for which a cleanup lock isn't immediately available; the next VACUUM will deal with them. If we're scanning the entire relation to advance relfrozenxid, we might need to wait, but only if there are tuples on the page that actually require freezing. These changes should greatly reduce the incidence of of vacuum processes getting "stuck". Simon Riggs and Robert Haas
2011-11-08 03:39:40 +01:00
{
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Processed as new/empty page (lock and pin released) */
Make VACUUM avoid waiting for a cleanup lock, where possible. In a regular VACUUM, it's OK to skip pages for which a cleanup lock isn't immediately available; the next VACUUM will deal with them. If we're scanning the entire relation to advance relfrozenxid, we might need to wait, but only if there are tuples on the page that actually require freezing. These changes should greatly reduce the incidence of of vacuum processes getting "stuck". Simon Riggs and Robert Haas
2011-11-08 03:39:40 +01:00
continue;
}
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Collect LP_DEAD items in dead_items array, count tuples */
if (lazy_scan_noprune(vacrel, buf, blkno, page, &hastup,
&recordfreespace))
Dept of second thoughts: the !scan_all exit mustn't increase scanned_pages. In the extreme edge case where contended pages are the only ones that escape being scanned, the previous commit would have allowed us to think that relfrozenxid could be advanced, which is exactly wrong.
2015-12-30 23:32:23 +01:00
{
Silence minor compiler warnings. Depending on compiler version and optimization level, we might get a complaint that lazy_scan_heap's "freespace" is used uninitialized. Compilers not aware that ereport(ERROR) doesn't return complained about bbsink_lz4_new(). Assigning "-1" to a uint64 value has unportable results; fortunately, the value of xlogreadsegno is unimportant when xlogreadfd is -1. (It looks to me like there is no need for xlogreadsegno to be static in the first place, but I didn't venture to change that.)
2022-02-13 19:06:55 +01:00
Size freespace = 0;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
Dept of second thoughts: the !scan_all exit mustn't increase scanned_pages. In the extreme edge case where contended pages are the only ones that escape being scanned, the previous commit would have allowed us to think that relfrozenxid could be advanced, which is exactly wrong.
2015-12-30 23:32:23 +01:00
/*
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
* Processed page successfully (without cleanup lock) -- just
* need to perform rel truncation and FSM steps, much like the
* lazy_scan_prune case. Don't bother trying to match its
* visibility map setting steps, though.
Dept of second thoughts: the !scan_all exit mustn't increase scanned_pages. In the extreme edge case where contended pages are the only ones that escape being scanned, the previous commit would have allowed us to think that relfrozenxid could be advanced, which is exactly wrong.
2015-12-30 23:32:23 +01:00
*/
if (hastup)
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->nonempty_pages = blkno + 1;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
if (recordfreespace)
freespace = PageGetHeapFreeSpace(page);
UnlockReleaseBuffer(buf);
if (recordfreespace)
RecordPageWithFreeSpace(vacrel->rel, blkno, freespace);
Dept of second thoughts: the !scan_all exit mustn't increase scanned_pages. In the extreme edge case where contended pages are the only ones that escape being scanned, the previous commit would have allowed us to think that relfrozenxid could be advanced, which is exactly wrong.
2015-12-30 23:32:23 +01:00
continue;
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Repair very-low-probability race condition between relation extension and VACUUM: in the interval between adding a new page to the relation and formatting it, it was possible for VACUUM to come along and decide it should format the page too. Though not harmful in itself, this would cause data loss if a third transaction were able to insert tuples into the vacuumed page before the original extender got control back.
2005-05-07 23:32:24 +02:00
/*
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
* lazy_scan_noprune could not do all required processing. Wait
* for a cleanup lock, and call lazy_scan_prune in the usual way.
Move page initialization from RelationAddExtraBlocks() to use. Previously we initialized pages when bulk extending in RelationAddExtraBlocks(). That has a major disadvantage: It ties RelationAddExtraBlocks() to heap, as other types of storage are likely to need different amounts of special space, have different amount of free space (previously determined by PageGetHeapFreeSpace()). That we're relying on initializing pages, but not WAL logging the initialization, also means the risk for getting "WARNING: relation \"%s\" page %u is uninitialized --- fixing" style warnings in vacuums after crashes/immediate shutdowns, is considerably higher. The warning sounds much more serious than what they are. Fix those two issues together by not initializing pages in RelationAddExtraPages() (but continue to do so in RelationGetBufferForTuple(), which is linked much more closely to heap), and accepting uninitialized pages as normal in vacuumlazy.c. When vacuumlazy encounters an empty page it now adds it to the FSM, but does nothing else. We chose to not issue a debug message, much less a warning in that case - it seems rarely useful, and quite likely to scare people unnecessarily. For now empty pages aren't added to the VM, because standbys would not re-discover such pages after a promotion. In contrast to other sources for empty pages, there's no corresponding WAL records triggering FSM updates during replay. Author: Andres Freund Reviewed-By: Tom Lane Discussion: https://postgr.es/m/20181219083945.6khtgm36mivonhva@alap3.anarazel.de
2019-01-28 22:15:11 +01:00
*/
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
Assert(vacrel->aggressive);
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockBufferForCleanup(buf);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Check for new or empty pages before lazy_scan_prune call */
if (lazy_scan_new_or_empty(vacrel, buf, blkno, page, false, vmbuffer))
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Processed as new/empty page (lock and pin released) */
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
continue;
}
HOT updates. When we update a tuple without changing any of its indexed columns, and the new version can be stored on the same heap page, we no longer generate extra index entries for the new version. Instead, index searches follow the HOT-chain links to ensure they find the correct tuple version. In addition, this patch introduces the ability to "prune" dead tuples on a per-page basis, without having to do a complete VACUUM pass to recover space. VACUUM is still needed to clean up dead index entries, however. Pavan Deolasee, with help from a bunch of other people.
2007-09-20 19:56:33 +02:00
/*
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
* Prune, freeze, and count tuples.
HOT updates. When we update a tuple without changing any of its indexed columns, and the new version can be stored on the same heap page, we no longer generate extra index entries for the new version. Instead, index searches follow the HOT-chain links to ensure they find the correct tuple version. In addition, this patch introduces the ability to "prune" dead tuples on a per-page basis, without having to do a complete VACUUM pass to recover space. VACUUM is still needed to clean up dead index entries, however. Pavan Deolasee, with help from a bunch of other people.
2007-09-20 19:56:33 +02:00
*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Accumulates details of remaining LP_DEAD line pointers on page in
vacuumlazy.c: fix remaining "dead tuple" references. Oversight in commit 4f8d9d12. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAD21AoDm38Em0bvRqeQKr4HPvOj65Y8cUgCP4idMk39iaLrxyw@mail.gmail.com
2021-11-30 20:40:33 +01:00
* dead_items array. This includes LP_DEAD line pointers that we
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* pruned ourselves, as well as existing LP_DEAD line pointers that
* were pruned some time earlier. Also considers freezing XIDs in the
* tuple headers of remaining items with storage.
HOT updates. When we update a tuple without changing any of its indexed columns, and the new version can be stored on the same heap page, we no longer generate extra index entries for the new version. Instead, index searches follow the HOT-chain links to ensure they find the correct tuple version. In addition, this patch introduces the ability to "prune" dead tuples on a per-page basis, without having to do a complete VACUUM pass to recover space. VACUUM is still needed to clean up dead index entries, however. Pavan Deolasee, with help from a bunch of other people.
2007-09-20 19:56:33 +02:00
*/
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
lazy_scan_prune(vacrel, buf, blkno, page, &prunestate);
pgindent run for 9.0, second run
2010-07-06 21:19:02 +02:00
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
Assert(!prunestate.all_visible || !prunestate.has_lpdead_items);
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/* Remember the location of the last page with nonremovable tuples */
if (prunestate.hastup)
vacrel->nonempty_pages = blkno + 1;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
if (vacrel->nindexes == 0)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
HOT updates. When we update a tuple without changing any of its indexed columns, and the new version can be stored on the same heap page, we no longer generate extra index entries for the new version. Instead, index searches follow the HOT-chain links to ensure they find the correct tuple version. In addition, this patch introduces the ability to "prune" dead tuples on a per-page basis, without having to do a complete VACUUM pass to recover space. VACUUM is still needed to clean up dead index entries, however. Pavan Deolasee, with help from a bunch of other people.
2007-09-20 19:56:33 +02:00
/*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Consider the need to do page-at-a-time heap vacuuming when
* using the one-pass strategy now.
Sync up our various ways of estimating pg_class.reltuples. VACUUM thought that reltuples represents the total number of tuples in the relation, while ANALYZE counted only live tuples. This can cause "flapping" in the value when background vacuums and analyzes happen separately. The planner's use of reltuples essentially assumes that it's the count of live (visible) tuples, so let's standardize on having it mean live tuples. Another issue is that the definition of "live tuple" isn't totally clear; what should be done with INSERT_IN_PROGRESS or DELETE_IN_PROGRESS tuples? ANALYZE's choices in this regard are made on the assumption that if the originating transaction commits at all, it will happen after ANALYZE finishes, so we should ignore the effects of the in-progress transaction --- unless it is our own transaction, and then we should count it. Let's propagate this definition into VACUUM, too. Likewise propagate this definition into CREATE INDEX, and into contrib/pgstattuple's pgstattuple_approx() function. Tomas Vondra, reviewed by Haribabu Kommi, some corrections by me Discussion: https://postgr.es/m/16db4468-edfa-830a-f921-39a50498e77e@2ndquadrant.com
2018-03-22 20:47:29 +01:00
*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* The one-pass strategy will never call lazy_vacuum(). The steps
* performed here can be thought of as the one-pass equivalent of
* a call to lazy_vacuum().
Sync up our various ways of estimating pg_class.reltuples. VACUUM thought that reltuples represents the total number of tuples in the relation, while ANALYZE counted only live tuples. This can cause "flapping" in the value when background vacuums and analyzes happen separately. The planner's use of reltuples essentially assumes that it's the count of live (visible) tuples, so let's standardize on having it mean live tuples. Another issue is that the definition of "live tuple" isn't totally clear; what should be done with INSERT_IN_PROGRESS or DELETE_IN_PROGRESS tuples? ANALYZE's choices in this regard are made on the assumption that if the originating transaction commits at all, it will happen after ANALYZE finishes, so we should ignore the effects of the in-progress transaction --- unless it is our own transaction, and then we should count it. Let's propagate this definition into VACUUM, too. Likewise propagate this definition into CREATE INDEX, and into contrib/pgstattuple's pgstattuple_approx() function. Tomas Vondra, reviewed by Haribabu Kommi, some corrections by me Discussion: https://postgr.es/m/16db4468-edfa-830a-f921-39a50498e77e@2ndquadrant.com
2018-03-22 20:47:29 +01:00
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
if (prunestate.has_lpdead_items)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
Size freespace;
Sync up our various ways of estimating pg_class.reltuples. VACUUM thought that reltuples represents the total number of tuples in the relation, while ANALYZE counted only live tuples. This can cause "flapping" in the value when background vacuums and analyzes happen separately. The planner's use of reltuples essentially assumes that it's the count of live (visible) tuples, so let's standardize on having it mean live tuples. Another issue is that the definition of "live tuple" isn't totally clear; what should be done with INSERT_IN_PROGRESS or DELETE_IN_PROGRESS tuples? ANALYZE's choices in this regard are made on the assumption that if the originating transaction commits at all, it will happen after ANALYZE finishes, so we should ignore the effects of the in-progress transaction --- unless it is our own transaction, and then we should count it. Let's propagate this definition into VACUUM, too. Likewise propagate this definition into CREATE INDEX, and into contrib/pgstattuple's pgstattuple_approx() function. Tomas Vondra, reviewed by Haribabu Kommi, some corrections by me Discussion: https://postgr.es/m/16db4468-edfa-830a-f921-39a50498e77e@2ndquadrant.com
2018-03-22 20:47:29 +01:00
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
lazy_vacuum_heap_page(vacrel, blkno, buf, 0, vmbuffer);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
/* Forget the LP_DEAD items that we just vacuumed */
dead_items->num_items = 0;
Improve vacuum code to track minimum Xids per table instead of per database. To this end, add a couple of columns to pg_class, relminxid and relvacuumxid, based on which we calculate the pg_database columns after each vacuum. We now force all databases to be vacuumed, even template ones. A backend noticing too old a database (meaning pg_database.datminxid is in danger of falling behind Xid wraparound) will signal the postmaster, which in turn will start an autovacuum iteration to process the offending database. In principle this is only there to cope with frozen (non-connectable) databases without forcing users to set them to connectable, but it could force regular user database to go through a database-wide vacuum at any time. Maybe we should warn users about this somehow. Of course the real solution will be to use autovacuum all the time ;-) There are some additional improvements we could have in this area: for example the vacuum code could be smarter about not updating pg_database for each table when called by autovacuum, and do it only once the whole autovacuum iteration is done. I updated the system catalogs documentation, but I didn't modify the maintenance section. Also having some regression tests for this would be nice but it's not really a very straightforward thing to do. Catalog version bumped due to system catalog changes.
2006-07-10 18:20:52 +02:00
/*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Periodically perform FSM vacuuming to make newly-freed
* space visible on upper FSM pages. Note we have not yet
* performed FSM processing for blkno.
Improve vacuum code to track minimum Xids per table instead of per database. To this end, add a couple of columns to pg_class, relminxid and relvacuumxid, based on which we calculate the pg_database columns after each vacuum. We now force all databases to be vacuumed, even template ones. A backend noticing too old a database (meaning pg_database.datminxid is in danger of falling behind Xid wraparound) will signal the postmaster, which in turn will start an autovacuum iteration to process the offending database. In principle this is only there to cope with frozen (non-connectable) databases without forcing users to set them to connectable, but it could force regular user database to go through a database-wide vacuum at any time. Maybe we should warn users about this somehow. Of course the real solution will be to use autovacuum all the time ;-) There are some additional improvements we could have in this area: for example the vacuum code could be smarter about not updating pg_database for each table when called by autovacuum, and do it only once the whole autovacuum iteration is done. I updated the system catalogs documentation, but I didn't modify the maintenance section. Also having some regression tests for this would be nice but it's not really a very straightforward thing to do. Catalog version bumped due to system catalog changes.
2006-07-10 18:20:52 +02:00
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
if (blkno - next_fsm_block_to_vacuum >= VACUUM_FSM_EVERY_PAGES)
{
FreeSpaceMapVacuumRange(vacrel->rel, next_fsm_block_to_vacuum,
blkno);
next_fsm_block_to_vacuum = blkno;
}
Fix recently-understood problems with handling of XID freezing, particularly in PITR scenarios. We now WAL-log the replacement of old XIDs with FrozenTransactionId, so that such replacement is guaranteed to propagate to PITR slave databases. Also, rather than relying on hint-bit updates to be preserved, pg_clog is not truncated until all instances of an XID are known to have been replaced by FrozenTransactionId. Add new GUC variables and pg_autovacuum columns to allow management of the freezing policy, so that users can trade off the size of pg_clog against the amount of freezing work done. Revise the already-existing code that forces autovacuum of tables approaching the wraparound point to make it more bulletproof; also, revise the autovacuum logic so that anti-wraparound vacuuming is done per-table rather than per-database. initdb forced because of changes in pg_class, pg_database, and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
2006-11-05 23:42:10 +01:00
Allow VACUUM to be run with index cleanup disabled. This commit adds a new reloption, vacuum_index_cleanup, which controls whether index cleanup is performed for a particular relation by default. It also adds a new option to the VACUUM command, INDEX_CLEANUP, which can be used to override the reloption. If neither the reloption nor the VACUUM option is used, the default is true, as before. Masahiko Sawada, reviewed and tested by Nathan Bossart, Alvaro Herrera, Kyotaro Horiguchi, Darafei Praliaskouski, and me. The wording of the documentation is mostly due to me. Discussion: http://postgr.es/m/CAD21AoAt5R3DNUZSjOoXDUY=naYPUOuffVsRzuTYMz29yLzQCA@mail.gmail.com
2019-04-04 20:58:53 +02:00
/*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Now perform FSM processing for blkno, and move on to next
* page.
Allow VACUUM to be run with index cleanup disabled. This commit adds a new reloption, vacuum_index_cleanup, which controls whether index cleanup is performed for a particular relation by default. It also adds a new option to the VACUUM command, INDEX_CLEANUP, which can be used to override the reloption. If neither the reloption nor the VACUUM option is used, the default is true, as before. Masahiko Sawada, reviewed and tested by Nathan Bossart, Alvaro Herrera, Kyotaro Horiguchi, Darafei Praliaskouski, and me. The wording of the documentation is mostly due to me. Discussion: http://postgr.es/m/CAD21AoAt5R3DNUZSjOoXDUY=naYPUOuffVsRzuTYMz29yLzQCA@mail.gmail.com
2019-04-04 20:58:53 +02:00
*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Our call to lazy_vacuum_heap_page() will have considered if
* it's possible to set all_visible/all_frozen independently
* of lazy_scan_prune(). Note that prunestate was invalidated
* by lazy_vacuum_heap_page() call.
Allow VACUUM to be run with index cleanup disabled. This commit adds a new reloption, vacuum_index_cleanup, which controls whether index cleanup is performed for a particular relation by default. It also adds a new option to the VACUUM command, INDEX_CLEANUP, which can be used to override the reloption. If neither the reloption nor the VACUUM option is used, the default is true, as before. Masahiko Sawada, reviewed and tested by Nathan Bossart, Alvaro Herrera, Kyotaro Horiguchi, Darafei Praliaskouski, and me. The wording of the documentation is mostly due to me. Discussion: http://postgr.es/m/CAD21AoAt5R3DNUZSjOoXDUY=naYPUOuffVsRzuTYMz29yLzQCA@mail.gmail.com
2019-04-04 20:58:53 +02:00
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
freespace = PageGetHeapFreeSpace(page);
pgindent run for 9.0, second run
2010-07-06 21:19:02 +02:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
UnlockReleaseBuffer(buf);
RecordPageWithFreeSpace(vacrel->rel, blkno, freespace);
continue;
}
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
/*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* There was no call to lazy_vacuum_heap_page() because pruning
* didn't encounter/create any LP_DEAD items that needed to be
* vacuumed. Prune state has not been invalidated, so proceed
* with prunestate-driven visibility map and FSM steps (just like
* the two-pass strategy).
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
*/
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
Assert(dead_items->num_items == 0);
Code review for patch to avoid second scan when vacuuming index-less table: avoid invoking LockBufferForCleanup without need, put out the same log message we would have before, minor code beautification.
2006-09-13 19:47:08 +02:00
}
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
* Handle setting visibility map bit based on information from the VM
* (as of last lazy_scan_skip() call), and from prunestate
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
if (!all_visible_according_to_vm && prunestate.all_visible)
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
{
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
uint8 flags = VISIBILITYMAP_ALL_VISIBLE;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
if (prunestate.all_frozen)
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
{
Assert(!TransactionIdIsValid(prunestate.visibility_cutoff_xid));
Department of second thoughts: remove PD_ALL_FROZEN. Commit a892234f830e832110f63fc0a2afce2fb21d1584 added a second bit per page to the visibility map, which still seems like a good idea, but it also added a second page-level bit alongside PD_ALL_VISIBLE to track whether the visibility map bit was set. That no longer seems like a clever plan, because we don't really need that bit for anything. We always clear both bits when the page is modified anyway. Patch by me, reviewed by Kyotaro Horiguchi and Masahiko Sawada.
2016-03-08 14:46:48 +01:00
flags |= VISIBILITYMAP_ALL_FROZEN;
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
}
Department of second thoughts: remove PD_ALL_FROZEN. Commit a892234f830e832110f63fc0a2afce2fb21d1584 added a second bit per page to the visibility map, which still seems like a good idea, but it also added a second page-level bit alongside PD_ALL_VISIBLE to track whether the visibility map bit was set. That no longer seems like a clever plan, because we don't really need that bit for anything. We always clear both bits when the page is modified anyway. Patch by me, reviewed by Kyotaro Horiguchi and Masahiko Sawada.
2016-03-08 14:46:48 +01:00
Ensure we MarkBufferDirty before visibilitymap_set() logs the heap page and sets the LSN. Otherwise a checkpoint could occur between those actions and leave us in an inconsistent state. Jeff Davis
2013-04-30 09:15:49 +02:00
/*
* It should never be the case that the visibility map page is set
* while the page-level bit is clear, but the reverse is allowed
Clarify some comments in vacuumlazy.c Author: Justin Pryzby Discussion: https://postgr.es/m/20200113004542.GA26045@telsasoft.com
2020-01-23 07:56:56 +01:00
* (if checksums are not enabled). Regardless, set both bits so
* that we get back in sync.
Ensure we MarkBufferDirty before visibilitymap_set() logs the heap page and sets the LSN. Otherwise a checkpoint could occur between those actions and leave us in an inconsistent state. Jeff Davis
2013-04-30 09:15:49 +02:00
*
* NB: If the heap page is all-visible but the VM bit is not set,
* we don't need to dirty the heap page. However, if checksums
* are enabled, we do need to make sure that the heap page is
* dirtied before passing it to visibilitymap_set(), because it
* may be logged. Given that this situation should only happen in
* rare cases after a crash, it is not worth optimizing.
*/
PageSetAllVisible(page);
MarkBufferDirty(buf);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
visibilitymap_set(vacrel->rel, blkno, buf, InvalidXLogRecPtr,
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
vmbuffer, prunestate.visibility_cutoff_xid,
flags);
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
}
/*
* As of PostgreSQL 9.2, the visibility map bit should never be set if
Fix more crash-safe visibility map bugs, and improve comments. In lazy_scan_heap, we could issue bogus warnings about incorrect information in the visibility map, because we checked the visibility map bit before locking the heap page, creating a race condition. Fix by rechecking the visibility map bit before we complain. Rejigger some related logic so that we rely on the possibly-outdated all_visible_according_to_vm value as little as possible. In heap_multi_insert, it's not safe to clear the visibility map bit before beginning the critical section. The visibility map is not crash-safe unless we treat clearing the bit as a critical operation. Specifically, if the transaction were to error out after we set the bit and before entering the critical section, we could end up writing the heap page to disk (with the bit cleared) and crashing before the visibility map page made it to disk. That would be bad. heap_insert has this correct, but somehow the order of operations got rearranged when heap_multi_insert was added. Also, add some more comments to visibilitymap_test, lazy_scan_heap, and IndexOnlyNext, expounding on concurrency issues. Per extensive code review by Andres Freund, and further review by Tom Lane, who also made the original report about the bogus warnings.
2012-06-07 18:25:41 +02:00
* the page-level bit is clear. However, it's possible that the bit
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
* got cleared after lazy_scan_skip() was called, so we must recheck
* with buffer lock before concluding that the VM is corrupt.
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
*/
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
else if (all_visible_according_to_vm && !PageIsAllVisible(page) &&
visibilitymap_get_status(vacrel->rel, blkno, &vmbuffer) != 0)
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
{
elog(WARNING, "page is not marked all-visible but visibility map bit is set in relation \"%s\" page %u",
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->relname, blkno);
visibilitymap_clear(vacrel->rel, blkno, vmbuffer,
Clear all-frozen visibilitymap status when locking tuples. Since a892234 & fd31cd265 the visibilitymap's freeze bit is used to avoid vacuuming the whole relation in anti-wraparound vacuums. Doing so correctly relies on not adding xids to the heap without also unsetting the visibilitymap flag. Tuple locking related code has not done so. To allow selectively resetting all-frozen - to avoid pessimizing heap_lock_tuple - allow to selectively reset the all-frozen with visibilitymap_clear(). To avoid having to use visibilitymap_get_status (e.g. via VM_ALL_FROZEN) inside a critical section, have visibilitymap_clear() return whether any bits have been reset. There's a remaining issue (denoted by XXX): After the PageIsAllVisible() check in heap_lock_tuple() and heap_lock_updated_tuple_rec() the page status could theoretically change. Practically that currently seems impossible, because updaters will hold a page level pin already. Due to the next beta coming up, it seems better to get the required WAL magic bump done before resolving this issue. The added flags field fields to xl_heap_lock and xl_heap_lock_updated require bumping the WAL magic. Since there's already been a catversion bump since the last beta, that's not an issue. Reviewed-By: Robert Haas, Amit Kapila and Andres Freund Author: Masahiko Sawada, heavily revised by Andres Freund Discussion: CAEepm=3fWAbWryVW9swHyLTY4sXVf0xbLvXqOwUoDiNCx9mBjQ@mail.gmail.com Backpatch: -
2016-07-18 11:01:13 +02:00
VISIBILITYMAP_VALID_BITS);
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
}
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
Don't throw a warning if vacuum sees PD_ALL_VISIBLE flag set on a page that contains newly-inserted tuples that according to our OldestXmin are not yet visible to everyone. The value returned by GetOldestXmin() is conservative, and it can move backwards on repeated calls, so if we see that contradiction between the PD_ALL_VISIBLE flag and status of tuples on the page, we have to assume it's because an earlier vacuum calculated a higher OldestXmin value, and all the tuples really are visible to everyone. We have received several reports of this bug, with the "PD_ALL_VISIBLE flag was incorrectly set in relation ..." warning appearing in logs. We were finally able to hunt it down with David Gould's help to run extra diagnostics in an environment where this happened frequently. Also reword the warning, per Robert Haas' suggestion, to not imply that the PD_ALL_VISIBLE flag is necessarily at fault, as it might also be a symptom of corruption on a tuple header. Backpatch to 8.4, where the PD_ALL_VISIBLE flag was introduced.
2011-03-08 19:13:52 +01:00
/*
snapshot scalability: Don't compute global horizons while building snapshots. To make GetSnapshotData() more scalable, it cannot not look at at each proc's xmin: While snapshot contents do not need to change whenever a read-only transaction commits or a snapshot is released, a proc's xmin is modified in those cases. The frequency of xmin modifications leads to, particularly on higher core count systems, many cache misses inside GetSnapshotData(), despite the data underlying a snapshot not changing. That is the most significant source of GetSnapshotData() scaling poorly on larger systems. Without accessing xmins, GetSnapshotData() cannot calculate accurate horizons / thresholds as it has so far. But we don't really have to: The horizons don't actually change that much between GetSnapshotData() calls. Nor are the horizons actually used every time a snapshot is built. The trick this commit introduces is to delay computation of accurate horizons until there use and using horizon boundaries to determine whether accurate horizons need to be computed. The use of RecentGlobal[Data]Xmin to decide whether a row version could be removed has been replaces with new GlobalVisTest* functions. These use two thresholds to determine whether a row can be pruned: 1) definitely_needed, indicating that rows deleted by XIDs >= definitely_needed are definitely still visible. 2) maybe_needed, indicating that rows deleted by XIDs < maybe_needed can definitely be removed GetSnapshotData() updates definitely_needed to be the xmin of the computed snapshot. When testing whether a row can be removed (with GlobalVisTestIsRemovableXid()) and the tested XID falls in between the two (i.e. XID >= maybe_needed && XID < definitely_needed) the boundaries can be recomputed to be more accurate. As it is not cheap to compute accurate boundaries, we limit the number of times that happens in short succession. As the boundaries used by GlobalVisTestIsRemovableXid() are never reset (with maybe_needed updated by GetSnapshotData()), it is likely that further test can benefit from an earlier computation of accurate horizons. To avoid regressing performance when old_snapshot_threshold is set (as that requires an accurate horizon to be computed), heap_page_prune_opt() doesn't unconditionally call TransactionIdLimitedForOldSnapshots() anymore. Both the computation of the limited horizon, and the triggering of errors (with SetOldSnapshotThresholdTimestamp()) is now only done when necessary to remove tuples. This commit just removes the accesses to PGXACT->xmin from GetSnapshotData(), but other members of PGXACT residing in the same cache line are accessed. Therefore this in itself does not result in a significant improvement. Subsequent commits will take advantage of the fact that GetSnapshotData() now does not need to access xmins anymore. Note: This contains a workaround in heap_page_prune_opt() to keep the snapshot_too_old tests working. While that workaround is ugly, the tests currently are not meaningful, and it seems best to address them separately. Author: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Thomas Munro <thomas.munro@gmail.com> Reviewed-By: David Rowley <dgrowleyml@gmail.com> Discussion: https://postgr.es/m/20200301083601.ews6hz5dduc3w2se@alap3.anarazel.de
2020-08-13 01:03:49 +02:00
* It's possible for the value returned by
* GetOldestNonRemovableTransactionId() to move backwards, so it's not
* wrong for us to see tuples that appear to not be visible to
* everyone yet, while PD_ALL_VISIBLE is already set. The real safe
* xmin value never moves backwards, but
* GetOldestNonRemovableTransactionId() is conservative and sometimes
* returns a value that's unnecessarily small, so if we see that
* contradiction it just means that the tuples that we think are not
* visible to everyone yet actually are, and the PD_ALL_VISIBLE flag
* is correct.
Don't throw a warning if vacuum sees PD_ALL_VISIBLE flag set on a page that contains newly-inserted tuples that according to our OldestXmin are not yet visible to everyone. The value returned by GetOldestXmin() is conservative, and it can move backwards on repeated calls, so if we see that contradiction between the PD_ALL_VISIBLE flag and status of tuples on the page, we have to assume it's because an earlier vacuum calculated a higher OldestXmin value, and all the tuples really are visible to everyone. We have received several reports of this bug, with the "PD_ALL_VISIBLE flag was incorrectly set in relation ..." warning appearing in logs. We were finally able to hunt it down with David Gould's help to run extra diagnostics in an environment where this happened frequently. Also reword the warning, per Robert Haas' suggestion, to not imply that the PD_ALL_VISIBLE flag is necessarily at fault, as it might also be a symptom of corruption on a tuple header. Backpatch to 8.4, where the PD_ALL_VISIBLE flag was introduced.
2011-03-08 19:13:52 +01:00
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* There should never be LP_DEAD items on a page with PD_ALL_VISIBLE
Don't throw a warning if vacuum sees PD_ALL_VISIBLE flag set on a page that contains newly-inserted tuples that according to our OldestXmin are not yet visible to everyone. The value returned by GetOldestXmin() is conservative, and it can move backwards on repeated calls, so if we see that contradiction between the PD_ALL_VISIBLE flag and status of tuples on the page, we have to assume it's because an earlier vacuum calculated a higher OldestXmin value, and all the tuples really are visible to everyone. We have received several reports of this bug, with the "PD_ALL_VISIBLE flag was incorrectly set in relation ..." warning appearing in logs. We were finally able to hunt it down with David Gould's help to run extra diagnostics in an environment where this happened frequently. Also reword the warning, per Robert Haas' suggestion, to not imply that the PD_ALL_VISIBLE flag is necessarily at fault, as it might also be a symptom of corruption on a tuple header. Backpatch to 8.4, where the PD_ALL_VISIBLE flag was introduced.
2011-03-08 19:13:52 +01:00
* set, however.
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
else if (prunestate.has_lpdead_items && PageIsAllVisible(page))
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
{
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
elog(WARNING, "page containing LP_DEAD items is marked as all-visible in relation \"%s\" page %u",
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->relname, blkno);
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
PageClearAllVisible(page);
Rearrange lazy_scan_heap to avoid visibility map race conditions. We must set the visibility map bit before releasing our exclusive lock on the heap page; otherwise, someone might clear the heap page bit before we set the visibility map bit, leading to a situation where the visibility map thinks the page is all-visible but it's really not. This problem has existed since 8.4, but it wasn't critical before we had index-only scans, since the worst case scenario was that the page wouldn't get vacuumed until the next scan_all vacuum. Along the way, a couple of minor, related improvements: (1) if we pause the heap scan to do an index vac cycle, release any visibility map page we're holding, since really long-running pins are not good for a variety of reasons; and (2) warn if we see a page that's marked all-visible in the visibility map but not on the page level, since that should never happen any more (it was allowed in previous releases, but not in 9.2).
2012-04-24 04:08:06 +02:00
MarkBufferDirty(buf);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
visibilitymap_clear(vacrel->rel, blkno, vmbuffer,
Clear all-frozen visibilitymap status when locking tuples. Since a892234 & fd31cd265 the visibilitymap's freeze bit is used to avoid vacuuming the whole relation in anti-wraparound vacuums. Doing so correctly relies on not adding xids to the heap without also unsetting the visibilitymap flag. Tuple locking related code has not done so. To allow selectively resetting all-frozen - to avoid pessimizing heap_lock_tuple - allow to selectively reset the all-frozen with visibilitymap_clear(). To avoid having to use visibilitymap_get_status (e.g. via VM_ALL_FROZEN) inside a critical section, have visibilitymap_clear() return whether any bits have been reset. There's a remaining issue (denoted by XXX): After the PageIsAllVisible() check in heap_lock_tuple() and heap_lock_updated_tuple_rec() the page status could theoretically change. Practically that currently seems impossible, because updaters will hold a page level pin already. Due to the next beta coming up, it seems better to get the required WAL magic bump done before resolving this issue. The added flags field fields to xl_heap_lock and xl_heap_lock_updated require bumping the WAL magic. Since there's already been a catversion bump since the last beta, that's not an issue. Reviewed-By: Robert Haas, Amit Kapila and Andres Freund Author: Masahiko Sawada, heavily revised by Andres Freund Discussion: CAEepm=3fWAbWryVW9swHyLTY4sXVf0xbLvXqOwUoDiNCx9mBjQ@mail.gmail.com Backpatch: -
2016-07-18 11:01:13 +02:00
VISIBILITYMAP_VALID_BITS);
Introduce visibility map. The visibility map is a bitmap with one bit per heap page, where a set bit indicates that all tuples on the page are visible to all transactions, and the page therefore doesn't need vacuuming. It is stored in a new relation fork. Lazy vacuum uses the visibility map to skip pages that don't need vacuuming. Vacuum is also responsible for setting the bits in the map. In the future, this can hopefully be used to implement index-only-scans, but we can't currently guarantee that the visibility map is always 100% up-to-date. In addition to the visibility map, there's a new PD_ALL_VISIBLE flag on each heap page, also indicating that all tuples on the page are visible to all transactions. It's important that this flag is kept up-to-date. It is also used to skip visibility tests in sequential scans, which gives a small performance gain on seqscans.
2008-12-03 14:05:22 +01:00
}
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
/*
Clarify some comments in vacuumlazy.c Author: Justin Pryzby Discussion: https://postgr.es/m/20200113004542.GA26045@telsasoft.com
2020-01-23 07:56:56 +01:00
* If the all-visible page is all-frozen but not marked as such yet,
* mark it as all-frozen. Note that all_frozen is only valid if
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
* all_visible is true, so we must check both prunestate fields.
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
else if (all_visible_according_to_vm && prunestate.all_visible &&
prunestate.all_frozen &&
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
!VM_ALL_FROZEN(vacrel->rel, blkno, &vmbuffer))
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
{
/*
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
* Avoid relying on all_visible_according_to_vm as a proxy for the
* page-level PD_ALL_VISIBLE bit being set, since it might have
* become stale -- even when all_visible is set in prunestate
*/
if (!PageIsAllVisible(page))
{
PageSetAllVisible(page);
MarkBufferDirty(buf);
}
/*
* Set the page all-frozen (and all-visible) in the VM.
*
* We can pass InvalidTransactionId as our visibility_cutoff_xid,
* since a snapshotConflictHorizon sufficient to make everything
* safe for REDO was logged when the page's tuples were frozen.
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
*/
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
Assert(!TransactionIdIsValid(prunestate.visibility_cutoff_xid));
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
visibilitymap_set(vacrel->rel, blkno, buf, InvalidXLogRecPtr,
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
vmbuffer, InvalidTransactionId,
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
VISIBILITYMAP_ALL_VISIBLE |
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
VISIBILITYMAP_ALL_FROZEN);
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
vacuumlazy.c: prefer the term "cleanup lock". The term "super-exclusive lock" is an acceptable synonym of "cleanup lock". Even still, switching from one term to the other in the same file is confusing. Standardize on "cleanup lock" within vacuumlazy.c. Per a complaint from Andres Freund.
2021-11-28 01:05:01 +01:00
* Final steps for block: drop cleanup lock, record free space in the
* FSM
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
if (prunestate.has_lpdead_items && vacrel->do_index_vacuuming)
{
/*
Truncate line pointer array during VACUUM. Teach VACUUM to truncate the line pointer array of each heap page when a contiguous group of LP_UNUSED line pointers appear at the end of the array -- these unused and unreferenced items are excluded. This process occurs during VACUUM's second pass over the heap, right after LP_DEAD line pointers on the page (those encountered/pruned during the first pass) are marked LP_UNUSED. Truncation avoids line pointer bloat with certain workloads, particularly those involving continual range DELETEs and bulk INSERTs against the same table. Also harden heapam code to check for an out-of-range page offset number in places where we weren't already doing so. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAH2-Wzn6a64PJM1Ggzm=uvx2otsopJMhFQj_g1rAj4GWr3ZSzw@mail.gmail.com
2021-04-07 17:47:15 +02:00
* Wait until lazy_vacuum_heap_rel() to save free space. This
* doesn't just save us some cycles; it also allows us to record
* any additional free space that lazy_vacuum_heap_page() will
* make available in cases where it's possible to truncate the
* page's line pointer array.
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
* Note: It's not in fact 100% certain that we really will call
* lazy_vacuum_heap_rel() -- lazy_vacuum() might yet opt to skip
* index vacuuming (and so must skip heap vacuuming). This is
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
* deemed okay because it only happens in emergencies, or when
* there is very little free space anyway. (Besides, we start
* recording free space in the FSM once index vacuuming has been
* abandoned.)
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
*
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* Note: The one-pass (no indexes) case is only supposed to make
* it this far when there were no LP_DEAD items during pruning.
*/
Assert(vacrel->nindexes > 0);
UnlockReleaseBuffer(buf);
}
else
{
Size freespace = PageGetHeapFreeSpace(page);
UnlockReleaseBuffer(buf);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
RecordPageWithFreeSpace(vacrel->rel, blkno, freespace);
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
vacrel->blkno = InvalidBlockNumber;
if (BufferIsValid(vmbuffer))
ReleaseBuffer(vmbuffer);
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/* report that everything is now scanned */
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
pgstat_progress_update_param(PROGRESS_VACUUM_HEAP_BLKS_SCANNED, blkno);
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
/* now we can compute the new value for pg_class.reltuples */
vacuumlazy.c: Standardize rel_pages terminology. VACUUM's rel_pages field indicates the size of the target heap rel just after the table_relation_vacuum() operation began. There are specific expectations around how rel_pages can be related to other nearby state. In particular, the range of rel_pages must contain every tuple in the relation whose tuple headers might contain an XID < OldestXmin. Consistently refer to the field as rel_pages to make this clearer and more discoverable. This is follow-up work to commit 73f6ec3d from earlier today. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20220311031351.sbge5m2bpvy2ttxg@alap3.anarazel.de
2022-03-12 22:20:45 +01:00
vacrel->new_live_tuples = vac_estimate_reltuples(vacrel->rel, rel_pages,
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
vacrel->scanned_pages,
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
vacrel->live_tuples);
Sync up our various ways of estimating pg_class.reltuples. VACUUM thought that reltuples represents the total number of tuples in the relation, while ANALYZE counted only live tuples. This can cause "flapping" in the value when background vacuums and analyzes happen separately. The planner's use of reltuples essentially assumes that it's the count of live (visible) tuples, so let's standardize on having it mean live tuples. Another issue is that the definition of "live tuple" isn't totally clear; what should be done with INSERT_IN_PROGRESS or DELETE_IN_PROGRESS tuples? ANALYZE's choices in this regard are made on the assumption that if the originating transaction commits at all, it will happen after ANALYZE finishes, so we should ignore the effects of the in-progress transaction --- unless it is our own transaction, and then we should count it. Let's propagate this definition into VACUUM, too. Likewise propagate this definition into CREATE INDEX, and into contrib/pgstattuple's pgstattuple_approx() function. Tomas Vondra, reviewed by Haribabu Kommi, some corrections by me Discussion: https://postgr.es/m/16db4468-edfa-830a-f921-39a50498e77e@2ndquadrant.com
2018-03-22 20:47:29 +01:00
Redefine pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Historically, we've considered the state with relpages and reltuples both zero as indicating that we do not know the table's tuple density. This is problematic because it's impossible to distinguish "never yet vacuumed" from "vacuumed and seen to be empty". In particular, a user cannot use VACUUM or ANALYZE to override the planner's normal heuristic that an empty table should not be believed to be empty because it is probably about to get populated. That heuristic is a good safety measure, so I don't care to abandon it, but there should be a way to override it if the table is indeed intended to stay empty. Hence, represent the initial state of ignorance by setting reltuples to -1 (relpages is still set to zero), and apply the minimum-ten-pages heuristic only when reltuples is still -1. If the table is empty, VACUUM or ANALYZE (but not CREATE INDEX) will override that to reltuples = relpages = 0, and then we'll plan on that basis. This requires a bunch of fiddly little changes, but we can get rid of some ugly kluges that were formerly needed to maintain the old definition. One notable point is that FDWs' GetForeignRelSize methods will see baserel->tuples = -1 when no ANALYZE has been done on the foreign table. That seems like a net improvement, since those methods were formerly also in the dark about what baserel->tuples = 0 really meant. Still, it is an API change. I bumped catversion because code predating this change would get confused by seeing reltuples = -1. Discussion: https://postgr.es/m/F02298E0-6EF4-49A1-BCB6-C484794D9ACC@thebuild.com
2020-08-30 18:21:51 +02:00
/*
* Also compute the total number of surviving heap entries. In the
* (unlikely) scenario that new_live_tuples is -1, take it as zero.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->new_rel_tuples =
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
Max(vacrel->new_live_tuples, 0) + vacrel->recently_dead_tuples +
vacrel->missed_dead_tuples;
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
Avoid holding vmbuffer pin after VACUUM. During VACUUM if we pause to perform a cycle of index cleanup we drop the vmbuffer pin, so we should do the same thing when heap scan completes. This avoids holding vmbuffer pin across the main index cleanup in VACUUM, which could be minutes or hours longer than necessary for correctness. Bug report and suggested fix from Pavan Deolasee
2012-12-03 19:53:31 +01:00
/*
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
* Do index vacuuming (call each index's ambulkdelete routine), then do
* related heap vacuuming
Avoid holding vmbuffer pin after VACUUM. During VACUUM if we pause to perform a cycle of index cleanup we drop the vmbuffer pin, so we should do the same thing when heap scan completes. This avoids holding vmbuffer pin across the main index cleanup in VACUUM, which could be minutes or hours longer than necessary for correctness. Bug report and suggested fix from Pavan Deolasee
2012-12-03 19:53:31 +01:00
*/
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
if (dead_items->num_items > 0)
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
lazy_vacuum(vacrel);
Clean up API for ambulkdelete/amvacuumcleanup as per today's discussion. This formulation requires every AM to provide amvacuumcleanup, unlike before, but it's surely a whole lot cleaner. Also, add an 'amstorage' column to pg_am so that we can get rid of hardwired knowledge in DefineOpClass().
2006-05-03 00:25:10 +02:00
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
/*
* Vacuum the remainder of the Free Space Map. We must do this whether or
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
* not there were indexes, and whether or not we bypassed index vacuuming.
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
*/
if (blkno > next_fsm_block_to_vacuum)
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
FreeSpaceMapVacuumRange(vacrel->rel, next_fsm_block_to_vacuum, blkno);
While vacuuming a large table, update upper-level FSM data every so often. VACUUM updates leaf-level FSM entries immediately after cleaning the corresponding heap blocks. fsmpage.c updates the intra-page search trees on the leaf-level FSM pages when this happens, but it does not touch the upper-level FSM pages, so that the released space might not actually be findable by searchers. Previously, updating the upper-level pages happened only at the conclusion of the VACUUM run, in a single FreeSpaceMapVacuum() call. This is bad because the VACUUM might get canceled before ever reaching that point, so that from the point of view of searchers no space has been freed at all, leading to table bloat. We can improve matters by updating the upper pages immediately after each cycle of index-cleaning and heap-cleaning, processing just the FSM pages corresponding to the range of heap blocks we have now fully cleaned. This adds a small amount of extra work, since the FSM pages leading down to each range boundary will be touched twice, but it's pretty negligible compared to everything else going on in a large VACUUM. If there are no indexes, VACUUM doesn't work in cycles but just cleans each heap page on first visit. In that case we just arbitrarily update upper FSM pages after each 8GB of heap. That maintains the goal of not letting all this work slide until the very end, and it doesn't seem worth expending extra complexity on a case that so seldom occurs in practice. In either case, the FSM is fully up to date before any attempt is made to truncate the relation, so that the most likely scenario for VACUUM cancellation no longer results in out-of-date upper FSM pages. When we do successfully truncate, adjusting the FSM to reflect that is now fully handled within FreeSpaceMapTruncateRel. Claudio Freire, reviewed by Masahiko Sawada and Jing Wang, some additional tweaks by me Discussion: https://postgr.es/m/CAGTBQpYR0uJCNTt3M5GOzBRHo+-GccNO1nCaQ8yEJmZKSW5q1A@mail.gmail.com
2018-03-29 17:29:54 +02:00
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
/* report all blocks vacuumed */
Fix problems in commit c16dc1aca5e01e6acaadfcf38f5fc964a381dc62. Vinayak Pokale provided a patch for a copy-and-paste error in a comment. I noticed that I'd use the word "automatically" nearby where I meant to talk about things being "atomic". Rahila Syed spotted a misplaced counter update. Fix all that stuff.
2016-03-16 18:54:04 +01:00
pgstat_progress_update_param(PROGRESS_VACUUM_HEAP_BLKS_VACUUMED, blkno);
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
vacuumlazy.c: Further consolidate resource allocation. Move remaining VACUUM resource allocation and deallocation code from lazy_scan_heap() to its caller, heap_vacuum_rel(). This finishes off work started by commit 73f6ec3d. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzk3fNBa_S3Ngi+16GQiyJ=AmUu3oUY99syMDTMRxitfyQ@mail.gmail.com
2022-04-04 20:53:33 +02:00
/* Do final index cleanup (call each index's amvacuumcleanup routine) */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
if (vacrel->nindexes > 0 && vacrel->do_index_cleanup)
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_cleanup_all_indexes(vacrel);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
/*
* lazy_scan_skip() -- set up range of skippable blocks using visibility map.
*
* lazy_scan_heap() calls here every time it needs to set up a new range of
* blocks to skip via the visibility map. Caller passes the next block in
* line. We return a next_unskippable_block for this range. When there are
* no skippable blocks we just return caller's next_block. The all-visible
* status of the returned block is set in *next_unskippable_allvis for caller,
* too. Block usually won't be all-visible (since it's unskippable), but it
* can be during aggressive VACUUMs (as well as in certain edge cases).
*
* Sets *skipping_current_range to indicate if caller should skip this range.
* Costs and benefits drive our decision. Very small ranges won't be skipped.
*
* Note: our opinion of which blocks can be skipped can go stale immediately.
* It's okay if caller "misses" a page whose all-visible or all-frozen marking
* was concurrently cleared, though. All that matters is that caller scan all
vacuumlazy.c: MultiXactIds are MXIDs, not XMIDs. Oversights in commits 0b018fab and f3c15cbe.
2022-04-21 03:29:02 +02:00
* pages whose tuples might contain XIDs < OldestXmin, or MXIDs < OldestMxact.
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
* (Actually, non-aggressive VACUUMs can choose to skip all-visible pages with
* older XIDs/MXIDs. The vacrel->skippedallvis flag will be set here when the
* choice to skip such a range is actually made, making everything safe.)
*/
static BlockNumber
lazy_scan_skip(LVRelState *vacrel, Buffer *vmbuffer, BlockNumber next_block,
bool *next_unskippable_allvis, bool *skipping_current_range)
{
BlockNumber rel_pages = vacrel->rel_pages,
next_unskippable_block = next_block,
nskippable_blocks = 0;
bool skipsallvis = false;
*next_unskippable_allvis = true;
while (next_unskippable_block < rel_pages)
{
uint8 mapbits = visibilitymap_get_status(vacrel->rel,
next_unskippable_block,
vmbuffer);
if ((mapbits & VISIBILITYMAP_ALL_VISIBLE) == 0)
{
Assert((mapbits & VISIBILITYMAP_ALL_FROZEN) == 0);
*next_unskippable_allvis = false;
break;
}
/*
* Caller must scan the last page to determine whether it has tuples
* (caller must have the opportunity to set vacrel->nonempty_pages).
* This rule avoids having lazy_truncate_heap() take access-exclusive
* lock on rel to attempt a truncation that fails anyway, just because
* there are tuples on the last page (it is likely that there will be
* tuples on other nearby pages as well, but those can be skipped).
*
* Implement this by always treating the last block as unsafe to skip.
*/
if (next_unskippable_block == rel_pages - 1)
break;
/* DISABLE_PAGE_SKIPPING makes all skipping unsafe */
if (!vacrel->skipwithvm)
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
{
/* Caller shouldn't rely on all_visible_according_to_vm */
*next_unskippable_allvis = false;
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
break;
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
}
Generalize how VACUUM skips all-frozen pages. Non-aggressive VACUUMs were at a gratuitous disadvantage (relative to aggressive VACUUMs) around advancing relfrozenxid and relminmxid before now. The issue only came up when concurrent activity unset some heap page's visibility map bit right as VACUUM was considering if the page should get counted in frozenskipped_pages. The non-aggressive case would recheck the all-frozen bit at this point. The aggressive case reasoned that the page (a skippable page) must have at least been all-frozen in the recent past, so skipping it won't make relfrozenxid advancement unsafe (which is never okay for aggressive VACUUMs). The recheck created a window for some other backend to confuse matters for VACUUM. If the page's VM bit turned out to be unset, VACUUM would conclude that the page was _never_ all-frozen. frozenskipped_pages was not incremented, and yet VACUUM couldn't back out of skipping at this late stage (it couldn't choose to scan the page instead). This made it unsafe to advance relfrozenxid later on. Consistently avoid the issue by generalizing how we skip frozen pages during aggressive VACUUMs: take the same approach when skipping any skippable page range during aggressive and non-aggressive VACUUMs alike. The new approach makes ranges (not individual pages) the fundamental unit of skipping using the visibility map. frozenskipped_pages is replaced with a boolean flag that represents whether some skippable range with one or more all-visible pages was actually skipped. It is safe for VACUUM to treat a page as all-frozen provided it at least had its all-frozen bit set after the OldestXmin cutoff was established. VACUUM is only required to scan pages that might have XIDs < OldestXmin (unfrozen XIDs) to be able to safely advance relfrozenxid. Tuples concurrently inserted on "skipped" pages can be thought of as equivalent to tuples concurrently inserted on a block >= rel_pages. It's possible that the issue this commit fixes hardly ever came up in practice. But we only had to be unlucky once to lose out on advancing relfrozenxid -- a single affected heap page was enough to throw VACUUM off. That seems like something to avoid on general principle. This is similar to an issue fixed by commit 44fa8488, which taught vacuumlazy.c to not give up on non-aggressive relfrozenxid advancement just because a cleanup lock wasn't immediately available on some heap page. Skipping an all-visible range is now explicitly structured as a choice made by non-aggressive VACUUMs, by weighing known costs (scanning extra skippable pages to freeze their tuples early) against known benefits (advancing relfrozenxid early). This works in essentially the same way as it always has (don't skip ranges < SKIP_PAGES_THRESHOLD). We could do much better here in the future by considering other relevant factors. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-Wzn6bGJGfOy3zSTJicKLw99PHJeSOQBOViKjSCinaxUKDQ@mail.gmail.com Discussion: https://postgr.es/m/CA%2BTgmoZiSOY6H7aadw5ZZGm7zYmfDzL6nwmL5V7GL4HgJgLF_w%40mail.gmail.com
2022-04-03 22:35:43 +02:00
/*
* Aggressive VACUUM caller can't skip pages just because they are
* all-visible. They may still skip all-frozen pages, which can't
* contain XIDs < OldestXmin (XIDs that aren't already frozen by now).
*/
if ((mapbits & VISIBILITYMAP_ALL_FROZEN) == 0)
{
if (vacrel->aggressive)
break;
/*
* All-visible block is safe to skip in non-aggressive case. But
* remember that the final range contains such a block for later.
*/
skipsallvis = true;
}
vacuum_delay_point();
next_unskippable_block++;
nskippable_blocks++;
}
/*
* We only skip a range with at least SKIP_PAGES_THRESHOLD consecutive
* pages. Since we're reading sequentially, the OS should be doing
* readahead for us, so there's no gain in skipping a page now and then.
* Skipping such a range might even discourage sequential detection.
*
* This test also enables more frequent relfrozenxid advancement during
* non-aggressive VACUUMs. If the range has any all-visible pages then
* skipping makes updating relfrozenxid unsafe, which is a real downside.
*/
if (nskippable_blocks < SKIP_PAGES_THRESHOLD)
*skipping_current_range = false;
else
{
*skipping_current_range = true;
if (skipsallvis)
vacrel->skippedallvis = true;
}
return next_unskippable_block;
}
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/*
* lazy_scan_new_or_empty() -- lazy_scan_heap() new/empty page handling.
*
* Must call here to handle both new and empty pages before calling
* lazy_scan_prune or lazy_scan_noprune, since they're not prepared to deal
* with new or empty pages.
*
* It's necessary to consider new pages as a special case, since the rules for
* maintaining the visibility map and FSM with empty pages are a little
Truncate line pointer array during heap pruning. Reclaim space from the line pointer array when heap pruning leaves behind a contiguous group of LP_UNUSED items at the end of the array. This happens during subsequent page defragmentation. Certain kinds of heap line pointer bloat are ameliorated by this new optimization. Follow-up work to commit 3c3b8a4b26, which taught VACUUM to truncate the line pointer array in about the same way during VACUUM's second pass over the heap. We now apply line pointer array truncation during both the first and the second pass over the heap made by VACUUM. We can also perform line pointer array truncation during opportunistic pruning. Matthias van de Meent, with small tweaks by me. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAEze2Wg36%2B4at2eWJNcYNiW2FJmht34x3YeX54ctUSs7kKoNcA%40mail.gmail.com
2022-04-08 00:42:12 +02:00
* different (though new pages can be truncated away during rel truncation).
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
*
* Empty pages are not really a special case -- they're just heap pages that
* have no allocated tuples (including even LP_UNUSED items). You might
* wonder why we need to handle them here all the same. It's only necessary
* because of a corner-case involving a hard crash during heap relation
* extension. If we ever make relation-extension crash safe, then it should
* no longer be necessary to deal with empty pages here (or new pages, for
* that matter).
*
* Caller must hold at least a shared lock. We might need to escalate the
* lock in that case, so the type of lock caller holds needs to be specified
* using 'sharelock' argument.
*
* Returns false in common case where caller should go on to call
* lazy_scan_prune (or lazy_scan_noprune). Otherwise returns true, indicating
* that lazy_scan_heap is done processing the page, releasing lock on caller's
* behalf.
*/
static bool
lazy_scan_new_or_empty(LVRelState *vacrel, Buffer buf, BlockNumber blkno,
Page page, bool sharelock, Buffer vmbuffer)
{
Size freespace;
if (PageIsNew(page))
{
/*
* All-zeroes pages can be left over if either a backend extends the
* relation by a single page, but crashes before the newly initialized
* page has been written out, or when bulk-extending the relation
* (which creates a number of empty pages at the tail end of the
* relation), and then enters them into the FSM.
*
* Note we do not enter the page into the visibilitymap. That has the
* downside that we repeatedly visit this page in subsequent vacuums,
* but otherwise we'll never discover the space on a promoted standby.
* The harm of repeated checking ought to normally not be too bad. The
* space usually should be used at some point, otherwise there
* wouldn't be any regular vacuums.
*
* Make sure these pages are in the FSM, to ensure they can be reused.
* Do that by testing if there's any space recorded for the page. If
* not, enter it. We do so after releasing the lock on the heap page,
* the FSM is approximate, after all.
*/
UnlockReleaseBuffer(buf);
if (GetRecordedFreeSpace(vacrel->rel, blkno) == 0)
{
freespace = BLCKSZ - SizeOfPageHeaderData;
RecordPageWithFreeSpace(vacrel->rel, blkno, freespace);
}
return true;
}
if (PageIsEmpty(page))
{
/*
* It seems likely that caller will always be able to get a cleanup
* lock on an empty page. But don't take any chances -- escalate to
* an exclusive lock (still don't need a cleanup lock, though).
*/
if (sharelock)
{
LockBuffer(buf, BUFFER_LOCK_UNLOCK);
LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
if (!PageIsEmpty(page))
{
/* page isn't new or empty -- keep lock and pin for now */
return false;
}
}
else
{
/* Already have a full cleanup lock (which is more than enough) */
}
/*
* Unlike new pages, empty pages are always set all-visible and
* all-frozen.
*/
if (!PageIsAllVisible(page))
{
START_CRIT_SECTION();
/* mark buffer dirty before writing a WAL record */
MarkBufferDirty(buf);
/*
* It's possible that another backend has extended the heap,
* initialized the page, and then failed to WAL-log the page due
* to an ERROR. Since heap extension is not WAL-logged, recovery
* might try to replay our record setting the page all-visible and
* find that the page isn't initialized, which will cause a PANIC.
* To prevent that, check whether the page has been previously
* WAL-logged, and if not, do that now.
*/
if (RelationNeedsWAL(vacrel->rel) &&
PageGetLSN(page) == InvalidXLogRecPtr)
log_newpage_buffer(buf, true);
PageSetAllVisible(page);
visibilitymap_set(vacrel->rel, blkno, buf, InvalidXLogRecPtr,
vmbuffer, InvalidTransactionId,
VISIBILITYMAP_ALL_VISIBLE | VISIBILITYMAP_ALL_FROZEN);
END_CRIT_SECTION();
}
freespace = PageGetHeapFreeSpace(page);
UnlockReleaseBuffer(buf);
RecordPageWithFreeSpace(vacrel->rel, blkno, freespace);
return true;
}
/* page isn't new or empty -- keep lock and pin */
return false;
}
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
* lazy_scan_prune() -- lazy_scan_heap() pruning and freezing.
*
* Caller must hold pin and buffer cleanup lock on the buffer.
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
*
* Prior to PostgreSQL 14 there were very rare cases where heap_page_prune()
* was allowed to disagree with our HeapTupleSatisfiesVacuum() call about
* whether or not a tuple should be considered DEAD. This happened when an
* inserting transaction concurrently aborted (after our heap_page_prune()
* call, before our HeapTupleSatisfiesVacuum() call). There was rather a lot
* of complexity just so we could deal with tuples that were DEAD to VACUUM,
* but nevertheless were left with storage after pruning.
*
* The approach we take now is to restart pruning when the race condition is
* detected. This allows heap_page_prune() to prune the tuples inserted by
* the now-aborted transaction. This is a little crude, but it guarantees
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* that any items that make it into the dead_items array are simple LP_DEAD
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
* line pointers, and that every remaining item with tuple storage is
* considered as a candidate for freezing.
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
static void
lazy_scan_prune(LVRelState *vacrel,
Buffer buf,
BlockNumber blkno,
Page page,
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
LVPagePruneState *prunestate)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
Relation rel = vacrel->rel;
OffsetNumber offnum,
maxoff;
ItemId itemid;
HeapTupleData tuple;
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
HTSV_Result res;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
int tuples_deleted,
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
tuples_frozen,
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
lpdead_items,
vacuumlazy.c: Remove obsolete num_tuples field. Commit 49c9d9fc unified VACUUM VERBOSE and autovacuum logging. It neglected to remove an old vacrel field that was only used by the old VACUUM VERBOSE, so remove it now. The previous num_tuples approach doesn't seem to have any real advantage over the approach VACUUM VERBOSE takes now (also the approach used by the autovacuum logging code), which is to show new_rel_tuples. new_rel_tuples is the possibly-estimated total number of tuples left in the table, whereas num_tuples meant the number of tuples encountered during the VACUUM operation, after pruning, without regard for tuples from pages skipped via the visibility map. In passing, reorder a related vacrel field for consistency.
2022-02-25 04:01:54 +01:00
live_tuples,
recently_dead_tuples;
Explain pruning pgstats accounting subtleties. Add a comment explaining why the pgstats accounting used during opportunistic heap pruning operations (to maintain the current number of dead tuples in the relation) needs to compensate by subtracting away the number of new LP_DEAD items. This is needed so it can avoid completely forgetting about tuples that become LP_DEAD items during pruning -- they should still count. It seems more natural to discuss this issue at the only relevant call site (opportunistic pruning), since the same issue does not apply to the only other caller (the VACUUM call site). Move everything there too. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzm7f+A6ej650gi_ifTgbhsadVW5cujAL3punpupHff5Yg@mail.gmail.com
2021-11-13 04:45:58 +01:00
int nnewlpdead;
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
HeapPageFreeze pagefrz;
int64 fpi_before = pgWalUsage.wal_fpi;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
OffsetNumber deadoffsets[MaxHeapTuplesPerPage];
Deduplicate freeze plans in freeze WAL records. Make heapam WAL records that describe freezing performed by VACUUM more space efficient by storing each distinct "freeze plan" once, alongside an array of associated page offset numbers (one per freeze plan). The freeze plans required for most heap pages tend to naturally have a great deal of redundancy, so this technique is very effective in practice. It often leads to freeze WAL records that are less than 20% of the size of equivalent WAL records generated using the previous approach. The freeze plan concept was introduced by commit 3b97e6823b, which fixed bugs in VACUUM's handling of MultiXacts. We retain the concept of freeze plans, but go back to using page offset number arrays. There is no loss of generality here because deduplication is an additive process that gets applied mechanically when FREEZE_PAGE WAL records are built. More than anything else, freeze plan deduplication is an optimization that reduces the marginal cost of freezing additional tuples on pages that will need to have at least one or two tuples frozen in any case. Ongoing work that adds page-level freezing to VACUUM will take full advantage of the improved cost profile through batching. Also refactor some of the details surrounding recovery conflicts needed to REDO freeze records in passing: make original execution responsible for generating a standard latestRemovedXid cutoff, rather than working backwards to get the same cutoff in the REDO routine. Bugfix commit 66fbcb0d2e did it the other way around, which is equivalent but obscures what's going on. Also rename the cutoff field from the WAL record/struct (rename the field cutoff_xid to latestRemovedXid to match similar WAL records). Processing of conflicts by REDO routines is already completely uniform, so tools like pg_waldump should present the information driving the process uniformly. There are two remaining WAL record types that still don't quite follow this convention (heapam's VISIBLE record type and SP-GiST's VACUUM_REDIRECT record type). They can be brought into line by later work that totally standardizes how the cutoffs are presented. Bump XLOG_PAGE_MAGIC. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Nathan Bossart <nathandbossart@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-Wz=XytErMnb8FAyFd+OQEbiipB0Q2FmFdXrggPL4VBnRYQ@mail.gmail.com
2022-11-15 16:48:41 +01:00
HeapTupleFreeze frozen[MaxHeapTuplesPerPage];
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
Assert(BufferGetBlockNumber(buf) == blkno);
Truncate line pointer array during heap pruning. Reclaim space from the line pointer array when heap pruning leaves behind a contiguous group of LP_UNUSED items at the end of the array. This happens during subsequent page defragmentation. Certain kinds of heap line pointer bloat are ameliorated by this new optimization. Follow-up work to commit 3c3b8a4b26, which taught VACUUM to truncate the line pointer array in about the same way during VACUUM's second pass over the heap. We now apply line pointer array truncation during both the first and the second pass over the heap made by VACUUM. We can also perform line pointer array truncation during opportunistic pruning. Matthias van de Meent, with small tweaks by me. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAEze2Wg36%2B4at2eWJNcYNiW2FJmht34x3YeX54ctUSs7kKoNcA%40mail.gmail.com
2022-04-08 00:42:12 +02:00
/*
* maxoff might be reduced following line pointer array truncation in
* heap_page_prune. That's safe for us to ignore, since the reclaimed
* space will continue to look like LP_UNUSED items below.
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
maxoff = PageGetMaxOffsetNumber(page);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
retry:
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
/* Initialize (or reset) page-level state */
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
pagefrz.freeze_required = false;
pagefrz.FreezePageRelfrozenXid = vacrel->NewRelfrozenXid;
pagefrz.FreezePageRelminMxid = vacrel->NewRelminMxid;
pagefrz.NoFreezePageRelfrozenXid = vacrel->NewRelfrozenXid;
pagefrz.NoFreezePageRelminMxid = vacrel->NewRelminMxid;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
tuples_deleted = 0;
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
tuples_frozen = 0;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
lpdead_items = 0;
live_tuples = 0;
vacuumlazy.c: Remove obsolete num_tuples field. Commit 49c9d9fc unified VACUUM VERBOSE and autovacuum logging. It neglected to remove an old vacrel field that was only used by the old VACUUM VERBOSE, so remove it now. The previous num_tuples approach doesn't seem to have any real advantage over the approach VACUUM VERBOSE takes now (also the approach used by the autovacuum logging code), which is to show new_rel_tuples. new_rel_tuples is the possibly-estimated total number of tuples left in the table, whereas num_tuples meant the number of tuples encountered during the VACUUM operation, after pruning, without regard for tuples from pages skipped via the visibility map. In passing, reorder a related vacrel field for consistency.
2022-02-25 04:01:54 +01:00
recently_dead_tuples = 0;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
* Prune all HOT-update chains in this page.
*
* We count tuples removed by the pruning step as tuples_deleted. Its
* final value can be thought of as the number of tuples that have been
* deleted from the table. It should not be confused with lpdead_items;
* lpdead_items's final value can be thought of as the number of tuples
* that were deleted from indexes.
*/
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
tuples_deleted = heap_page_prune(rel, buf, vacrel->vistest,
Explain pruning pgstats accounting subtleties. Add a comment explaining why the pgstats accounting used during opportunistic heap pruning operations (to maintain the current number of dead tuples in the relation) needs to compensate by subtracting away the number of new LP_DEAD items. This is needed so it can avoid completely forgetting about tuples that become LP_DEAD items during pruning -- they should still count. It seems more natural to discuss this issue at the only relevant call site (opportunistic pruning), since the same issue does not apply to the only other caller (the VACUUM call site). Move everything there too. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-Wzm7f+A6ej650gi_ifTgbhsadVW5cujAL3punpupHff5Yg@mail.gmail.com
2021-11-13 04:45:58 +01:00
InvalidTransactionId, 0, &nnewlpdead,
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
&vacrel->offnum);
/*
* Now scan the page to collect LP_DEAD items and check for tuples
* requiring freezing among remaining tuples with storage
*/
prunestate->hastup = false;
prunestate->has_lpdead_items = false;
prunestate->all_visible = true;
prunestate->all_frozen = true;
prunestate->visibility_cutoff_xid = InvalidTransactionId;
for (offnum = FirstOffsetNumber;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
bool totally_frozen;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
* Set the offset number so that we can display it along with any
* error that occurred while processing this tuple.
*/
vacrel->offnum = offnum;
itemid = PageGetItemId(page, offnum);
if (!ItemIdIsUsed(itemid))
continue;
/* Redirect items mustn't be touched */
if (ItemIdIsRedirected(itemid))
{
Adjust VACUUM hastup LP_REDIRECT comments. The term "truncation" has been ambiguous since commit 10a8d13823 added line pointer array truncation during heap pruning. Clear things up by specifying that we're talking about rel truncation here, to match nearby comments that apply to tuples with storage.
2023-01-02 19:18:22 +01:00
/* page makes rel truncation unsafe */
prunestate->hastup = true;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
continue;
}
if (ItemIdIsDead(itemid))
{
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
/*
* Deliberately don't set hastup for LP_DEAD items. We make the
* soft assumption that any LP_DEAD items encountered here will
* become LP_UNUSED later on, before count_nondeletable_pages is
* reached. If we don't make this assumption then rel truncation
* will only happen every other VACUUM, at most. Besides, VACUUM
* must treat hastup/nonempty_pages as provisional no matter how
* LP_DEAD items are handled (handled here, or handled later on).
*
* Also deliberately delay unsetting all_visible until just before
* we return to lazy_scan_heap caller, as explained in full below.
* (This is another case where it's useful to anticipate that any
* LP_DEAD items will become LP_UNUSED during the ongoing VACUUM.)
*/
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
deadoffsets[lpdead_items++] = offnum;
continue;
}
Assert(ItemIdIsNormal(itemid));
ItemPointerSet(&(tuple.t_self), blkno, offnum);
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid);
tuple.t_len = ItemIdGetLength(itemid);
tuple.t_tableOid = RelationGetRelid(rel);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
/*
* DEAD tuples are almost always pruned into LP_DEAD line pointers by
* heap_page_prune(), but it's possible that the tuple state changed
* since heap_page_prune() looked. Handle that here by restarting.
* (See comments at the top of function for a full explanation.)
*/
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
res = HeapTupleSatisfiesVacuum(&tuple, vacrel->cutoffs.OldestXmin,
buf);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
if (unlikely(res == HEAPTUPLE_DEAD))
goto retry;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
* The criteria for counting a tuple as live in this block need to
* match what analyze.c's acquire_sample_rows() does, otherwise VACUUM
* and ANALYZE may produce wildly different reltuples values, e.g.
* when there are many recently-dead tuples.
*
* The logic here is a bit simpler than acquire_sample_rows(), as
* VACUUM can't run inside a transaction block, which makes some cases
* impossible (e.g. in-progress insert from the same transaction).
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* We treat LP_DEAD items (which are the closest thing to DEAD tuples
* that might be seen here) differently, too: we assume that they'll
* become LP_UNUSED before VACUUM finishes. This difference is only
* superficial. VACUUM effectively agrees with ANALYZE about DEAD
* items, in the end. VACUUM won't remember LP_DEAD items, but only
* because they're not supposed to be left behind when it is done.
* (Cases where we bypass index vacuuming will violate this optimistic
* assumption, but the overall impact of that should be negligible.)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
switch (res)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
case HEAPTUPLE_LIVE:
/*
* Count it as live. Not only is this natural, but it's also
* what acquire_sample_rows() does.
*/
live_tuples++;
/*
* Is the tuple definitely visible to all transactions?
*
* NB: Like with per-tuple hint bits, we can't set the
* PD_ALL_VISIBLE flag if the inserter committed
* asynchronously. See SetHintBits for more info. Check that
* the tuple is hinted xmin-committed because of that.
*/
if (prunestate->all_visible)
{
TransactionId xmin;
if (!HeapTupleHeaderXminCommitted(tuple.t_data))
{
prunestate->all_visible = false;
break;
}
/*
* The inserter definitely committed. But is it old enough
* that everyone sees it as committed?
*/
xmin = HeapTupleHeaderGetXmin(tuple.t_data);
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
if (!TransactionIdPrecedes(xmin,
vacrel->cutoffs.OldestXmin))
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
prunestate->all_visible = false;
break;
}
/* Track newest xmin on page. */
Avoid special XID snapshotConflictHorizon values. Don't allow VACUUM to WAL-log the value FrozenTransactionId as the snapshotConflictHorizon of freezing or visibility map related WAL records. The only special XID value that's an allowable snapshotConflictHorizon is InvalidTransactionId, which is interpreted as "record definitely doesn't require a recovery conflict". Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-02 19:16:51 +01:00
if (TransactionIdFollows(xmin, prunestate->visibility_cutoff_xid) &&
TransactionIdIsNormal(xmin))
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
prunestate->visibility_cutoff_xid = xmin;
}
break;
case HEAPTUPLE_RECENTLY_DEAD:
/*
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
* If tuple is recently dead then we must not remove it from
* the relation. (We only remove items that are LP_DEAD from
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
* pruning.)
*/
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
recently_dead_tuples++;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
prunestate->all_visible = false;
break;
case HEAPTUPLE_INSERT_IN_PROGRESS:
/*
* We do not count these rows as live, because we expect the
* inserting transaction to update the counters at commit, and
* we assume that will happen only after we report our
* results. This assumption is a bit shaky, but it is what
* acquire_sample_rows() does, so be consistent.
*/
prunestate->all_visible = false;
break;
case HEAPTUPLE_DELETE_IN_PROGRESS:
/* This is an expected case during concurrent vacuum */
prunestate->all_visible = false;
/*
* Count such rows as live. As above, we assume the deleting
* transaction will commit and update the counters after we
* report.
*/
live_tuples++;
break;
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
Truncate line pointer array during heap pruning. Reclaim space from the line pointer array when heap pruning leaves behind a contiguous group of LP_UNUSED items at the end of the array. This happens during subsequent page defragmentation. Certain kinds of heap line pointer bloat are ameliorated by this new optimization. Follow-up work to commit 3c3b8a4b26, which taught VACUUM to truncate the line pointer array in about the same way during VACUUM's second pass over the heap. We now apply line pointer array truncation during both the first and the second pass over the heap made by VACUUM. We can also perform line pointer array truncation during opportunistic pruning. Matthias van de Meent, with small tweaks by me. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAEze2Wg36%2B4at2eWJNcYNiW2FJmht34x3YeX54ctUSs7kKoNcA%40mail.gmail.com
2022-04-08 00:42:12 +02:00
prunestate->hastup = true; /* page makes rel truncation unsafe */
Deduplicate freeze plans in freeze WAL records. Make heapam WAL records that describe freezing performed by VACUUM more space efficient by storing each distinct "freeze plan" once, alongside an array of associated page offset numbers (one per freeze plan). The freeze plans required for most heap pages tend to naturally have a great deal of redundancy, so this technique is very effective in practice. It often leads to freeze WAL records that are less than 20% of the size of equivalent WAL records generated using the previous approach. The freeze plan concept was introduced by commit 3b97e6823b, which fixed bugs in VACUUM's handling of MultiXacts. We retain the concept of freeze plans, but go back to using page offset number arrays. There is no loss of generality here because deduplication is an additive process that gets applied mechanically when FREEZE_PAGE WAL records are built. More than anything else, freeze plan deduplication is an optimization that reduces the marginal cost of freezing additional tuples on pages that will need to have at least one or two tuples frozen in any case. Ongoing work that adds page-level freezing to VACUUM will take full advantage of the improved cost profile through batching. Also refactor some of the details surrounding recovery conflicts needed to REDO freeze records in passing: make original execution responsible for generating a standard latestRemovedXid cutoff, rather than working backwards to get the same cutoff in the REDO routine. Bugfix commit 66fbcb0d2e did it the other way around, which is equivalent but obscures what's going on. Also rename the cutoff field from the WAL record/struct (rename the field cutoff_xid to latestRemovedXid to match similar WAL records). Processing of conflicts by REDO routines is already completely uniform, so tools like pg_waldump should present the information driving the process uniformly. There are two remaining WAL record types that still don't quite follow this convention (heapam's VISIBLE record type and SP-GiST's VACUUM_REDIRECT record type). They can be brought into line by later work that totally standardizes how the cutoffs are presented. Bump XLOG_PAGE_MAGIC. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Nathan Bossart <nathandbossart@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-Wz=XytErMnb8FAyFd+OQEbiipB0Q2FmFdXrggPL4VBnRYQ@mail.gmail.com
2022-11-15 16:48:41 +01:00
/* Tuple with storage -- consider need to freeze */
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
if (heap_prepare_freeze_tuple(tuple.t_data, &vacrel->cutoffs, &pagefrz,
&frozen[tuples_frozen], &totally_frozen))
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
Deduplicate freeze plans in freeze WAL records. Make heapam WAL records that describe freezing performed by VACUUM more space efficient by storing each distinct "freeze plan" once, alongside an array of associated page offset numbers (one per freeze plan). The freeze plans required for most heap pages tend to naturally have a great deal of redundancy, so this technique is very effective in practice. It often leads to freeze WAL records that are less than 20% of the size of equivalent WAL records generated using the previous approach. The freeze plan concept was introduced by commit 3b97e6823b, which fixed bugs in VACUUM's handling of MultiXacts. We retain the concept of freeze plans, but go back to using page offset number arrays. There is no loss of generality here because deduplication is an additive process that gets applied mechanically when FREEZE_PAGE WAL records are built. More than anything else, freeze plan deduplication is an optimization that reduces the marginal cost of freezing additional tuples on pages that will need to have at least one or two tuples frozen in any case. Ongoing work that adds page-level freezing to VACUUM will take full advantage of the improved cost profile through batching. Also refactor some of the details surrounding recovery conflicts needed to REDO freeze records in passing: make original execution responsible for generating a standard latestRemovedXid cutoff, rather than working backwards to get the same cutoff in the REDO routine. Bugfix commit 66fbcb0d2e did it the other way around, which is equivalent but obscures what's going on. Also rename the cutoff field from the WAL record/struct (rename the field cutoff_xid to latestRemovedXid to match similar WAL records). Processing of conflicts by REDO routines is already completely uniform, so tools like pg_waldump should present the information driving the process uniformly. There are two remaining WAL record types that still don't quite follow this convention (heapam's VISIBLE record type and SP-GiST's VACUUM_REDIRECT record type). They can be brought into line by later work that totally standardizes how the cutoffs are presented. Bump XLOG_PAGE_MAGIC. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Nathan Bossart <nathandbossart@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-Wz=XytErMnb8FAyFd+OQEbiipB0Q2FmFdXrggPL4VBnRYQ@mail.gmail.com
2022-11-15 16:48:41 +01:00
/* Save prepared freeze plan for later */
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
frozen[tuples_frozen++].offset = offnum;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
/*
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
* If any tuple isn't either totally frozen already or eligible to
* become totally frozen (according to its freeze plan), then the page
* definitely cannot be set all-frozen in the visibility map later on
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
*/
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
if (!totally_frozen)
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
prunestate->all_frozen = false;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
/*
* We have now divided every item on the page into either an LP_DEAD item
* that will need to be vacuumed in indexes later, or a LP_NORMAL tuple
* that remains and needs to be considered for freezing now (LP_UNUSED and
* LP_REDIRECT items also remain, but are of no further interest to us).
*/
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
vacrel->offnum = InvalidOffsetNumber;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
* Freeze the page when heap_prepare_freeze_tuple indicates that at least
* one XID/MXID from before FreezeLimit/MultiXactCutoff is present. Also
* freeze when pruning generated an FPI, if doing so means that we set the
* page all-frozen afterwards (might not happen until final heap pass).
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
if (pagefrz.freeze_required || tuples_frozen == 0 ||
(prunestate->all_visible && prunestate->all_frozen &&
Revert "Add eager and lazy freezing strategies to VACUUM." This reverts commit 4d417992613949af35530b4e8e83670c4e67e1b2. Broad concerns about regressions caused by eager freezing strategy have been raised. Whether or not these concerns can be worked through in any time frame is far from certain. Discussion: https://postgr.es/m/20230126004347.gepcmyenk2csxrri@awork3.anarazel.de
2023-01-26 07:22:27 +01:00
fpi_before != pgWalUsage.wal_fpi))
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
/*
* We're freezing the page. Our final NewRelfrozenXid doesn't need to
* be affected by the XIDs that are just about to be frozen anyway.
*/
vacrel->NewRelfrozenXid = pagefrz.FreezePageRelfrozenXid;
vacrel->NewRelminMxid = pagefrz.FreezePageRelminMxid;
if (tuples_frozen == 0)
{
/*
Refine the definition of page-level freezing. Improve comments added by commit 1de58df4 which describe the lazy_scan_prune "freeze the page" path. These newly revised comments are based on suggestions from Jeff Davis. In passing, remove nearby visibility_cutoff_xid comments left over from commit 6daeeb1f. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/ebc857107fe3edd422ef8a65191ca4a8da568b9b.camel@j-davis.com
2023-01-03 19:08:55 +01:00
* We have no freeze plans to execute, so there's no added cost
Pre-beta mechanical code beautification. Run pgindent, pgperltidy, and reformat-dat-files. This set of diffs is a bit larger than typical. We've updated to pg_bsd_indent 2.1.2, which properly indents variable declarations that have multi-line initialization expressions (the continuation lines are now indented one tab stop). We've also updated to perltidy version 20230309 and changed some of its settings, which reduces its desire to add whitespace to lines to make assignments etc. line up. Going forward, that should make for fewer random-seeming changes to existing code. Discussion: https://postgr.es/m/20230428092545.qfb3y5wcu4cm75ur@alvherre.pgsql
2023-05-19 23:24:48 +02:00
* from following the freeze path. That's why it was chosen. This
* is important in the case where the page only contains totally
* frozen tuples at this point (perhaps only following pruning).
* Such pages can be marked all-frozen in the VM by our caller,
* even though none of its tuples were newly frozen here (note
* that the "no freeze" path never sets pages all-frozen).
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
*
* We never increment the frozen_pages instrumentation counter
* here, since it only counts pages with newly frozen tuples
* (don't confuse that with pages newly set all-frozen in VM).
*/
}
else
{
TransactionId snapshotConflictHorizon;
vacrel->frozen_pages++;
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
/*
* We can use visibility_cutoff_xid as our cutoff for conflicts
* when the whole page is eligible to become all-frozen in the VM
* once we're done with it. Otherwise we generate a conservative
* cutoff by stepping back from OldestXmin.
*/
if (prunestate->all_visible && prunestate->all_frozen)
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
{
/* Using same cutoff when setting VM is now unnecessary */
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
snapshotConflictHorizon = prunestate->visibility_cutoff_xid;
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
prunestate->visibility_cutoff_xid = InvalidTransactionId;
}
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
else
{
/* Avoids false conflicts when hot_standby_feedback in use */
snapshotConflictHorizon = vacrel->cutoffs.OldestXmin;
TransactionIdRetreat(snapshotConflictHorizon);
}
/* Execute all freeze plans for page as a single atomic action */
heap_freeze_execute_prepared(vacrel->rel, buf,
snapshotConflictHorizon,
frozen, tuples_frozen);
}
}
else
{
/*
* Page requires "no freeze" processing. It might be set all-visible
* in the visibility map, but it can never be set all-frozen.
*/
vacrel->NewRelfrozenXid = pagefrz.NoFreezePageRelfrozenXid;
vacrel->NewRelminMxid = pagefrz.NoFreezePageRelminMxid;
prunestate->all_frozen = false;
tuples_frozen = 0; /* avoid miscounts in instrumentation */
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
/*
* VACUUM will call heap_page_is_all_visible() during the second pass over
* the heap to determine all_visible and all_frozen for the page -- this
* is a specialized version of the logic from this function. Now that
* we've finished pruning and freezing, make sure that we're in total
* agreement with heap_page_is_all_visible() using an assertion.
*/
#ifdef USE_ASSERT_CHECKING
/* Note that all_frozen value does not matter when !all_visible */
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
if (prunestate->all_visible && lpdead_items == 0)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
TransactionId cutoff;
bool all_frozen;
if (!heap_page_is_all_visible(vacrel, buf, &cutoff, &all_frozen))
Assert(false);
Avoid special XID snapshotConflictHorizon values. Don't allow VACUUM to WAL-log the value FrozenTransactionId as the snapshotConflictHorizon of freezing or visibility map related WAL records. The only special XID value that's an allowable snapshotConflictHorizon is InvalidTransactionId, which is interpreted as "record definitely doesn't require a recovery conflict". Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-02 19:16:51 +01:00
Assert(!TransactionIdIsValid(cutoff) ||
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
cutoff == prunestate->visibility_cutoff_xid);
}
#endif
/*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* Now save details of the LP_DEAD items from the page in vacrel
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
if (lpdead_items > 0)
{
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
VacDeadItems *dead_items = vacrel->dead_items;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
ItemPointerData tmp;
vacrel->lpdead_item_pages++;
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
prunestate->has_lpdead_items = true;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
ItemPointerSetBlockNumber(&tmp, blkno);
for (int i = 0; i < lpdead_items; i++)
{
ItemPointerSetOffsetNumber(&tmp, deadoffsets[i]);
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
dead_items->items[dead_items->num_items++] = tmp;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
Assert(dead_items->num_items <= dead_items->max_items);
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
pgstat_progress_update_param(PROGRESS_VACUUM_NUM_DEAD_TUPLES,
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
dead_items->num_items);
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
/*
* It was convenient to ignore LP_DEAD items in all_visible earlier on
* to make the choice of whether or not to freeze the page unaffected
* by the short-term presence of LP_DEAD items. These LP_DEAD items
* were effectively assumed to be LP_UNUSED items in the making. It
* doesn't matter which heap pass (initial pass or final pass) ends up
* setting the page all-frozen, as long as the ongoing VACUUM does it.
*
* Now that freezing has been finalized, unset all_visible. It needs
* to reflect the present state of things, as expected by our caller.
*/
prunestate->all_visible = false;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
/* Finally, add page-local counts to whole-VACUUM counts */
vacrel->tuples_deleted += tuples_deleted;
Instrument freezing in autovacuum log reports. Add a new line to log reports from autovacuum (as well as VACUUM VERBOSE output) that shows information about freezing. Emphasis is placed on the total number of heap pages that had one or more tuples frozen by VACUUM. The total number of tuples frozen is also shown. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Janes <jeff.janes@gmail.com> Discussion: https://postgr.es/m/CAH2-WznTY6D0zyE8VLrC6Gd4kh_HGAXxnTPtcOQOOsxzLx9zog@mail.gmail.com
2022-09-08 19:29:39 +02:00
vacrel->tuples_frozen += tuples_frozen;
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
vacrel->lpdead_items += lpdead_items;
vacrel->live_tuples += live_tuples;
vacuumlazy.c: Remove obsolete num_tuples field. Commit 49c9d9fc unified VACUUM VERBOSE and autovacuum logging. It neglected to remove an old vacrel field that was only used by the old VACUUM VERBOSE, so remove it now. The previous num_tuples approach doesn't seem to have any real advantage over the approach VACUUM VERBOSE takes now (also the approach used by the autovacuum logging code), which is to show new_rel_tuples. new_rel_tuples is the possibly-estimated total number of tuples left in the table, whereas num_tuples meant the number of tuples encountered during the VACUUM operation, after pruning, without regard for tuples from pages skipped via the visibility map. In passing, reorder a related vacrel field for consistency.
2022-02-25 04:01:54 +01:00
vacrel->recently_dead_tuples += recently_dead_tuples;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/*
* lazy_scan_noprune() -- lazy_scan_prune() without pruning or freezing
*
* Caller need only hold a pin and share lock on the buffer, unlike
* lazy_scan_prune, which requires a full cleanup lock. While pruning isn't
* performed here, it's quite possible that an earlier opportunistic pruning
* operation left LP_DEAD items behind. We'll at least collect any such items
* in the dead_items array for removal from indexes.
*
* For aggressive VACUUM callers, we may return false to indicate that a full
* cleanup lock is required for processing by lazy_scan_prune. This is only
* necessary when the aggressive VACUUM needs to freeze some tuple XIDs from
* one or more tuples on the page. We always return true for non-aggressive
* callers.
*
* See lazy_scan_prune for an explanation of hastup return flag.
* recordfreespace flag instructs caller on whether or not it should do
* generic FSM processing for page.
*/
static bool
lazy_scan_noprune(LVRelState *vacrel,
Buffer buf,
BlockNumber blkno,
Page page,
bool *hastup,
bool *recordfreespace)
{
OffsetNumber offnum,
maxoff;
int lpdead_items,
live_tuples,
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
recently_dead_tuples,
missed_dead_tuples;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
HeapTupleHeader tupleheader;
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
TransactionId NoFreezePageRelfrozenXid = vacrel->NewRelfrozenXid;
MultiXactId NoFreezePageRelminMxid = vacrel->NewRelminMxid;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
OffsetNumber deadoffsets[MaxHeapTuplesPerPage];
Assert(BufferGetBlockNumber(buf) == blkno);
*hastup = false; /* for now */
*recordfreespace = false; /* for now */
lpdead_items = 0;
live_tuples = 0;
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
recently_dead_tuples = 0;
missed_dead_tuples = 0;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
ItemId itemid;
HeapTupleData tuple;
vacrel->offnum = offnum;
itemid = PageGetItemId(page, offnum);
if (!ItemIdIsUsed(itemid))
continue;
if (ItemIdIsRedirected(itemid))
{
*hastup = true;
continue;
}
if (ItemIdIsDead(itemid))
{
/*
* Deliberately don't set hastup=true here. See same point in
* lazy_scan_prune for an explanation.
*/
deadoffsets[lpdead_items++] = offnum;
continue;
}
*hastup = true; /* page prevents rel truncation */
tupleheader = (HeapTupleHeader) PageGetItem(page, itemid);
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
if (heap_tuple_should_freeze(tupleheader, &vacrel->cutoffs,
&NoFreezePageRelfrozenXid,
&NoFreezePageRelminMxid))
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
{
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
/* Tuple with XID < FreezeLimit (or MXID < MultiXactCutoff) */
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
if (vacrel->aggressive)
{
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
/*
* Aggressive VACUUMs must always be able to advance rel's
* relfrozenxid to a value >= FreezeLimit (and be able to
* advance rel's relminmxid to a value >= MultiXactCutoff).
* The ongoing aggressive VACUUM won't be able to do that
vacuumlazy.c: MultiXactIds are MXIDs, not XMIDs. Oversights in commits 0b018fab and f3c15cbe.
2022-04-21 03:29:02 +02:00
* unless it can freeze an XID (or MXID) from this tuple now.
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
*
* The only safe option is to have caller perform processing
* of this page using lazy_scan_prune. Caller might have to
* wait a while for a cleanup lock, but it can't be helped.
*/
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
vacrel->offnum = InvalidOffsetNumber;
return false;
}
/*
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
* Non-aggressive VACUUMs are under no obligation to advance
* relfrozenxid (even by one XID). We can be much laxer here.
*
* Currently we always just accept an older final relfrozenxid
* and/or relminmxid value. We never make caller wait or work a
* little harder, even when it likely makes sense to do so.
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
*/
}
ItemPointerSet(&(tuple.t_self), blkno, offnum);
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid);
tuple.t_len = ItemIdGetLength(itemid);
tuple.t_tableOid = RelationGetRelid(vacrel->rel);
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
switch (HeapTupleSatisfiesVacuum(&tuple, vacrel->cutoffs.OldestXmin,
buf))
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
{
case HEAPTUPLE_DELETE_IN_PROGRESS:
case HEAPTUPLE_LIVE:
/*
* Count both cases as live, just like lazy_scan_prune
*/
live_tuples++;
break;
case HEAPTUPLE_DEAD:
/*
* There is some useful work for pruning to do, that won't be
* done due to failure to get a cleanup lock.
*/
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
missed_dead_tuples++;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
break;
case HEAPTUPLE_RECENTLY_DEAD:
/*
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
* Count in recently_dead_tuples, just like lazy_scan_prune
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
*/
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
recently_dead_tuples++;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
break;
case HEAPTUPLE_INSERT_IN_PROGRESS:
/*
* Do not count these rows as live, just like lazy_scan_prune
*/
break;
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
}
vacrel->offnum = InvalidOffsetNumber;
/*
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
* By here we know for sure that caller can put off freezing and pruning
* this particular page until the next VACUUM. Remember its details now.
* (lazy_scan_prune expects a clean slate, so we have to do this last.)
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
*/
Add page-level freezing to VACUUM. Teach VACUUM to decide on whether or not to trigger freezing at the level of whole heap pages. Individual XIDs and MXIDs fields from tuple headers now trigger freezing of whole pages, rather than independently triggering freezing of each individual tuple header field. Managing the cost of freezing over time now significantly influences when and how VACUUM freezes. The overall amount of WAL written is the single most important freezing related cost, in general. Freezing each page's tuples together in batch allows VACUUM to take full advantage of the freeze plan WAL deduplication optimization added by commit 9e540599. Also teach VACUUM to trigger page-level freezing whenever it detects that heap pruning generated an FPI. We'll have already written a large amount of WAL just to do that much, so it's very likely a good idea to get freezing out of the way for the page early. This only happens in cases where it will directly lead to marking the page all-frozen in the visibility map. In most cases "freezing a page" removes all XIDs < OldestXmin, and all MXIDs < OldestMxact. It doesn't quite work that way in certain rare cases involving MultiXacts, though. It is convenient to define "freeze the page" in a way that gives FreezeMultiXactId the leeway to put off the work of processing an individual tuple's xmax whenever it happens to be a MultiXactId that would require an expensive second pass to process aggressively (allocating a new multi is especially worth avoiding here). FreezeMultiXactId is eager when processing is cheap (as it usually is), and lazy in the event of an individual multi that happens to require expensive second pass processing. This avoids regressions related to processing of multis that page-level freezing might otherwise cause. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzkFok_6EAHuK39GaW4FjEFQsY=3J0AAd6FXk93u-Xq3Fg@mail.gmail.com
2022-12-28 17:50:47 +01:00
vacrel->NewRelfrozenXid = NoFreezePageRelfrozenXid;
vacrel->NewRelminMxid = NoFreezePageRelminMxid;
Set relfrozenxid to oldest extant XID seen by VACUUM. When VACUUM set relfrozenxid before now, it set it to whatever value was used to determine which tuples to freeze -- the FreezeLimit cutoff. This approach was very naive. The relfrozenxid invariant only requires that new relfrozenxid values be <= the oldest extant XID remaining in the table (at the point that the VACUUM operation ends), which in general might be much more recent than FreezeLimit. VACUUM now carefully tracks the oldest remaining XID/MultiXactId as it goes (the oldest remaining values _after_ lazy_scan_prune processing). The final values are set as the table's new relfrozenxid and new relminmxid in pg_class at the end of each VACUUM. The oldest XID might come from a tuple's xmin, xmax, or xvac fields. It might even come from one of the table's remaining MultiXacts. Final relfrozenxid values must still be >= FreezeLimit in an aggressive VACUUM (FreezeLimit still acts as a lower bound on the final value that aggressive VACUUM can set relfrozenxid to). Since standard VACUUMs still make no guarantees about advancing relfrozenxid, they might as well set relfrozenxid to a value from well before FreezeLimit when the opportunity presents itself. In general standard VACUUMs may now set relfrozenxid to any value > the original relfrozenxid and <= OldestXmin. Credit for the general idea of using the oldest extant XID to set pg_class.relfrozenxid at the end of VACUUM goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkymFbz6D_vL+jmqSn_5q1wsFvFrE+37yLgL_Rkfd6Gzg@mail.gmail.com
2022-04-03 18:57:21 +02:00
/* Save any LP_DEAD items found on the page in dead_items array */
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
if (vacrel->nindexes == 0)
{
/* Using one-pass strategy (since table has no indexes) */
if (lpdead_items > 0)
{
/*
* Perfunctory handling for the corner case where a single pass
* strategy VACUUM cannot get a cleanup lock, and it turns out
* that there is one or more LP_DEAD items: just count the LP_DEAD
* items as missed_dead_tuples instead. (This is a bit dishonest,
* but it beats having to maintain specialized heap vacuuming code
* forever, for vanishingly little benefit.)
*/
*hastup = true;
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
missed_dead_tuples += lpdead_items;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
}
*recordfreespace = true;
}
else if (lpdead_items == 0)
{
/*
* Won't be vacuuming this page later, so record page's freespace in
* the FSM now
*/
*recordfreespace = true;
}
else
{
VacDeadItems *dead_items = vacrel->dead_items;
ItemPointerData tmp;
/*
* Page has LP_DEAD items, and so any references/TIDs that remain in
* indexes will be deleted during index vacuuming (and then marked
* LP_UNUSED in the heap)
*/
vacrel->lpdead_item_pages++;
ItemPointerSetBlockNumber(&tmp, blkno);
for (int i = 0; i < lpdead_items; i++)
{
ItemPointerSetOffsetNumber(&tmp, deadoffsets[i]);
dead_items->items[dead_items->num_items++] = tmp;
}
Assert(dead_items->num_items <= dead_items->max_items);
pgstat_progress_update_param(PROGRESS_VACUUM_NUM_DEAD_TUPLES,
dead_items->num_items);
vacrel->lpdead_items += lpdead_items;
/*
* Assume that we'll go on to vacuum this heap page during final pass
* over the heap. Don't record free space until then.
*/
*recordfreespace = false;
}
/*
* Finally, add relevant page-local counts to whole-VACUUM counts
*/
vacuumlazy.c: Remove obsolete num_tuples field. Commit 49c9d9fc unified VACUUM VERBOSE and autovacuum logging. It neglected to remove an old vacrel field that was only used by the old VACUUM VERBOSE, so remove it now. The previous num_tuples approach doesn't seem to have any real advantage over the approach VACUUM VERBOSE takes now (also the approach used by the autovacuum logging code), which is to show new_rel_tuples. new_rel_tuples is the possibly-estimated total number of tuples left in the table, whereas num_tuples meant the number of tuples encountered during the VACUUM operation, after pruning, without regard for tuples from pages skipped via the visibility map. In passing, reorder a related vacrel field for consistency.
2022-02-25 04:01:54 +01:00
vacrel->live_tuples += live_tuples;
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
vacrel->recently_dead_tuples += recently_dead_tuples;
vacrel->missed_dead_tuples += missed_dead_tuples;
if (missed_dead_tuples > 0)
vacrel->missed_dead_pages++;
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
/* Caller won't need to call lazy_scan_prune with same page */
return true;
}
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
vacuumlazy.c: fix "garbage tuples" reference. Another minor oversight in commit 4f8d9d12.
2022-01-12 23:13:35 +01:00
* Main entry point for index vacuuming and heap vacuuming.
*
* Removes items collected in dead_items from table's indexes, then marks the
* same items LP_UNUSED in the heap. See the comments above lazy_scan_heap
* for full details.
*
* Also empties dead_items, freeing up space for later TIDs.
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
*
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
* We may choose to bypass index vacuuming at this point, though only when the
* ongoing VACUUM operation will definitely only have one index scan/round of
vacuumlazy.c: Remove obsolete 'onecall' comment. Remove obsolete reference to lazy_vacuum()'s onecall argument. The function argument was removed by commit 3499df0dee. Also remove adjoining comment block that introduces the wraparound failsafe concept. Talking about the failsafe here no longer makes sense, since lazy_vacuum() (and related functions) are no longer the only place where the failsafe might be triggered. This has been the case since commit c242baa4a8 taught VACUUM to consider triggering the failsafe mechanism during its initial heap scan.
2021-09-25 19:22:53 +02:00
* index vacuuming.
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
static void
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
lazy_vacuum(LVRelState *vacrel)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
{
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
bool bypass;
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/* Should not end up here with no indexes */
Assert(vacrel->nindexes > 0);
Assert(vacrel->lpdead_item_pages > 0);
if (!vacrel->do_index_vacuuming)
{
Assert(!vacrel->do_index_cleanup);
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
vacrel->dead_items->num_items = 0;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
return;
}
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
/*
* Consider bypassing index vacuuming (and heap vacuuming) entirely.
*
* We currently only do this in cases where the number of LP_DEAD items
* for the entire VACUUM operation is close to zero. This avoids sharp
* discontinuities in the duration and overhead of successive VACUUM
* operations that run against the same table with a fixed workload.
* Ideally, successive VACUUM operations will behave as if there are
* exactly zero LP_DEAD items in cases where there are close to zero.
*
* This is likely to be helpful with a table that is continually affected
* by UPDATEs that can mostly apply the HOT optimization, but occasionally
* have small aberrations that lead to just a few heap pages retaining
* only one or two LP_DEAD items. This is pretty common; even when the
* DBA goes out of their way to make UPDATEs use HOT, it is practically
* impossible to predict whether HOT will be applied in 100% of cases.
* It's far easier to ensure that 99%+ of all UPDATEs against a table use
* HOT through careful tuning.
*/
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
bypass = false;
if (vacrel->consider_bypass_optimization && vacrel->rel_pages > 0)
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
{
BlockNumber threshold;
Assert(vacrel->num_index_scans == 0);
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
Assert(vacrel->lpdead_items == vacrel->dead_items->num_items);
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
Assert(vacrel->do_index_vacuuming);
Assert(vacrel->do_index_cleanup);
/*
* This crossover point at which we'll start to do index vacuuming is
* expressed as a percentage of the total number of heap pages in the
* table that are known to have at least one LP_DEAD item. This is
* much more important than the total number of LP_DEAD items, since
* it's a proxy for the number of heap pages whose visibility map bits
* cannot be set on account of bypassing index and heap vacuuming.
*
* We apply one further precautionary test: the space currently used
* to store the TIDs (TIDs that now all point to LP_DEAD items) must
* not exceed 32MB. This limits the risk that we will bypass index
* vacuuming again and again until eventually there is a VACUUM whose
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* dead_items space is not CPU cache resident.
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
*
* We don't take any special steps to remember the LP_DEAD items (such
pgstat: stats collector references in comments. Soon the stats collector will be no more, with statistics instead getting stored in shared memory. There are a lot of references to the stats collector in comments. This commit replaces most of these references with "cumulative statistics system", with the remaining ones getting replaced as part of subsequent commits. This is done separately from the - quite large - shared memory statistics patch to make review easier. Author: Andres Freund <andres@anarazel.de> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Reviewed-By: Thomas Munro <thomas.munro@gmail.com> Reviewed-By: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Discussion: https://postgr.es/m/20220303021600.hs34ghqcw6zcokdh@alap3.anarazel.de Discussion: https://postgr.es/m/20220308205351.2xcn6k4x5yivcxyd@alap3.anarazel.de
2022-04-06 22:56:06 +02:00
* as counting them in our final update to the stats system) when the
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
* optimization is applied. Though the accounting used in analyze.c's
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
* acquire_sample_rows() will recognize the same LP_DEAD items as dead
pgstat: stats collector references in comments. Soon the stats collector will be no more, with statistics instead getting stored in shared memory. There are a lot of references to the stats collector in comments. This commit replaces most of these references with "cumulative statistics system", with the remaining ones getting replaced as part of subsequent commits. This is done separately from the - quite large - shared memory statistics patch to make review easier. Author: Andres Freund <andres@anarazel.de> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Reviewed-By: Thomas Munro <thomas.munro@gmail.com> Reviewed-By: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Discussion: https://postgr.es/m/20220303021600.hs34ghqcw6zcokdh@alap3.anarazel.de Discussion: https://postgr.es/m/20220308205351.2xcn6k4x5yivcxyd@alap3.anarazel.de
2022-04-06 22:56:06 +02:00
* rows in its own stats report, that's okay. The discrepancy should
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
* be negligible. If this optimization is ever expanded to cover more
* cases then this may need to be reconsidered.
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
*/
threshold = (double) vacrel->rel_pages * BYPASS_THRESHOLD_PAGES;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
bypass = (vacrel->lpdead_item_pages < threshold &&
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
vacrel->lpdead_items < MAXDEADITEMS(32L * 1024L * 1024L));
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
}
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
if (bypass)
Teach VACUUM to bypass unnecessary index vacuuming. VACUUM has never needed to call ambulkdelete() for each index in cases where there are precisely zero TIDs in its dead_tuples array by the end of its first pass over the heap (also its only pass over the heap in this scenario). Index vacuuming is simply not required when this happens. Index cleanup will still go ahead, but in practice most calls to amvacuumcleanup() are usually no-ops when there were zero preceding ambulkdelete() calls. In short, VACUUM has generally managed to avoid index scans when there were clearly no index tuples to delete from indexes. But cases with _close to_ no index tuples to delete were another matter -- a round of ambulkdelete() calls took place (one per index), each of which performed a full index scan. VACUUM now behaves just as if there were zero index tuples to delete in cases where there are in fact "virtually zero" such tuples. That is, it can now bypass index vacuuming and heap vacuuming as an optimization (though not index cleanup). Whether or not VACUUM bypasses indexes is determined dynamically, based on the just-observed number of heap pages in the table that have one or more LP_DEAD items (LP_DEAD items in heap pages have a 1:1 correspondence with index tuples that still need to be deleted from each index in the worst case). We only skip index vacuuming when 2% or less of the table's pages have one or more LP_DEAD items -- bypassing index vacuuming as an optimization must not noticeably impede setting bits in the visibility map. As a further condition, the dead_tuples array (i.e. VACUUM's array of LP_DEAD item TIDs) must not exceed 32MB at the point that the first pass over the heap finishes, which is also when the decision to bypass is made. (The VACUUM must also have been able to fit all TIDs in its maintenance_work_mem-bound dead_tuples space, though with a default maintenance_work_mem setting it can't matter.) This avoids surprising jumps in the duration and overhead of routine vacuuming with workloads where successive VACUUM operations consistently have almost zero dead index tuples. The number of LP_DEAD items may well accumulate over multiple VACUUM operations, before finally the threshold is crossed and VACUUM performs conventional index vacuuming. Even then, the optimization will have avoided a great deal of largely unnecessary index vacuuming. In the future we may teach VACUUM to skip index vacuuming on a per-index basis, using a much more sophisticated approach. For now we only consider the extreme cases, where we can be quite confident that index vacuuming just isn't worth it using simple heuristics. Also log information about how many heap pages have one or more LP_DEAD items when autovacuum logging is enabled. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmkebqPd4MVGuPTOS9bMFvp9MDs5cRTCOsv1rQJ3jCbXw@mail.gmail.com
2021-04-08 01:14:54 +02:00
{
/*
* There are almost zero TIDs. Behave as if there were precisely
* zero: bypass index vacuuming, but do index cleanup.
*
* We expect that the ongoing VACUUM operation will finish very
* quickly, so there is no point in considering speeding up as a
* failsafe against wraparound failure. (Index cleanup is expected to
* finish very quickly in cases where there were no ambulkdelete()
* calls.)
*/
vacrel->do_index_vacuuming = false;
}
else if (lazy_vacuum_all_indexes(vacrel))
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
{
/*
* We successfully completed a round of index vacuuming. Do related
* heap vacuuming now.
*/
lazy_vacuum_heap_rel(vacrel);
}
else
{
/*
* Failsafe case.
*
Update high level vacuumlazy.c comments. Update vacuumlazy.c file header comments (as well as comments above the lazy_scan_heap function) that were largely written before the introduction of the HOT optimization, when lazy_scan_heap did far less, and didn't actually prune during its initial heap pass. Since lazy_scan_heap now outsources far more work to lower level functions, it makes sense to introduce the function by talking about the high level invariant that dictates the order in which each phase takes place. Also deemphasize the case where we run out of memory for TIDs, since delaying that discussion makes it easier to talk about issues of central importance. Finally, remove discussion of parallel VACUUM from header comments. These don't add much, and are in the wrong place.
2021-11-27 23:29:43 +01:00
* We attempted index vacuuming, but didn't finish a full round/full
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
* index scan. This happens when relfrozenxid or relminmxid is too
* far in the past.
*
* From this point on the VACUUM operation will do no further index
Don't truncate heap when VACUUM's failsafe is in effect. It seems like a good idea to bypass heap truncation when the wraparound failsafe mechanism (which was added in commit 1e55e7d1) is in effect. Deliberately don't bypass heap truncation in the INDEX_CLEANUP=off case, even though it is similar to the failsafe case. There is already a separate reloption (and related VACUUM parameter) for that. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAD21AoDWRh6oTN5T8wa+cpZUVpHXET8BJ8Da7WHVHpwkPP6KLg@mail.gmail.com
2021-04-13 21:58:31 +02:00
* vacuuming or heap vacuuming. This VACUUM operation won't end up
* back here again.
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
*/
Make vacuum failsafe_active globally visible While vacuuming a table in failsafe mode, VacuumCostActive should not be re-enabled. This currently isn't a problem because vacuum cost parameters are only refreshed in between vacuuming tables and failsafe status is reset for every table. In preparation for allowing vacuum cost parameters to be updated more frequently, elevate LVRelState->failsafe_active to a global, VacuumFailsafeActive, which will be checked when determining whether or not to re-enable vacuum cost-related delays. Author: Melanie Plageman <melanieplageman@gmail.com> Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Daniel Gustafsson <daniel@yesql.se> Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Reviewed-by: Robert Haas <robertmhaas@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/CAAKRu_ZngzqnEODc7LmS1NH04Kt6Y9huSjz5pp7%2BDXhrjDA0gw%40mail.gmail.com
2023-04-07 00:54:08 +02:00
Assert(VacuumFailsafeActive);
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
}
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
/*
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
* Forget the LP_DEAD items that we just vacuumed (or just decided to not
* vacuum)
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
*/
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
vacrel->dead_items->num_items = 0;
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
}
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
/*
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
* lazy_vacuum_all_indexes() -- Main entry for index vacuuming
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
*
* Returns true in the common case when all indexes were successfully
* vacuumed. Returns false in rare cases where we determined that the ongoing
* VACUUM operation is at risk of taking too long to finish, leading to
* wraparound failure.
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
*/
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
static bool
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_vacuum_all_indexes(LVRelState *vacrel)
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
{
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
bool allindexes = true;
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
double old_live_tuples = vacrel->rel->rd_rel->reltuples;
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Assert(vacrel->nindexes > 0);
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
Assert(vacrel->do_index_vacuuming);
Assert(vacrel->do_index_cleanup);
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
/* Precheck for XID wraparound emergencies */
if (lazy_check_wraparound_failsafe(vacrel))
{
/* Wraparound emergency -- don't even start an index scan */
return false;
}
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
/* Report that we are now vacuuming indexes */
pgstat_progress_update_param(PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_PHASE_VACUUM_INDEX);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (!ParallelVacuumIsActive(vacrel))
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
for (int idx = 0; idx < vacrel->nindexes; idx++)
{
Relation indrel = vacrel->indrels[idx];
IndexBulkDeleteResult *istat = vacrel->indstats[idx];
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
vacrel->indstats[idx] = lazy_vacuum_one_index(indrel, istat,
old_live_tuples,
vacrel);
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
if (lazy_check_wraparound_failsafe(vacrel))
{
/* Wraparound emergency -- end current index scan */
allindexes = false;
break;
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
}
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
}
else
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/* Outsource everything to parallel variant */
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
parallel_vacuum_bulkdel_all_indexes(vacrel->pvs, old_live_tuples,
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
vacrel->num_index_scans);
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
/*
* Do a postcheck to consider applying wraparound failsafe now. Note
* that parallel VACUUM only gets the precheck and this postcheck.
*/
if (lazy_check_wraparound_failsafe(vacrel))
allindexes = false;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
}
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
/*
* We delete all LP_DEAD items from the first heap pass in all indexes on
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
* each call here (except calls where we choose to do the failsafe). This
* makes the next call to lazy_vacuum_heap_rel() safe (except in the event
* of the failsafe triggering, which prevents the next call from taking
* place).
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
*/
Assert(vacrel->num_index_scans > 0 ||
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
vacrel->dead_items->num_items == vacrel->lpdead_items);
Make vacuum failsafe_active globally visible While vacuuming a table in failsafe mode, VacuumCostActive should not be re-enabled. This currently isn't a problem because vacuum cost parameters are only refreshed in between vacuuming tables and failsafe status is reset for every table. In preparation for allowing vacuum cost parameters to be updated more frequently, elevate LVRelState->failsafe_active to a global, VacuumFailsafeActive, which will be checked when determining whether or not to re-enable vacuum cost-related delays. Author: Melanie Plageman <melanieplageman@gmail.com> Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Daniel Gustafsson <daniel@yesql.se> Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Reviewed-by: Robert Haas <robertmhaas@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/CAAKRu_ZngzqnEODc7LmS1NH04Kt6Y9huSjz5pp7%2BDXhrjDA0gw%40mail.gmail.com
2023-04-07 00:54:08 +02:00
Assert(allindexes || VacuumFailsafeActive);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
/*
* Increase and report the number of index scans.
*
* We deliberately include the case where we started a round of bulk
* deletes that we weren't able to finish due to the failsafe triggering.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->num_index_scans++;
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
pgstat_progress_update_param(PROGRESS_VACUUM_NUM_INDEX_VACUUMS,
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->num_index_scans);
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
return allindexes;
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
* lazy_vacuum_heap_rel() -- second pass over the heap for two pass strategy
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* This routine marks LP_DEAD items in vacrel->dead_items array as LP_UNUSED.
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
* Pages that never had lazy_scan_prune record LP_DEAD items are not visited
* at all.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
Truncate line pointer array during VACUUM. Teach VACUUM to truncate the line pointer array of each heap page when a contiguous group of LP_UNUSED line pointers appear at the end of the array -- these unused and unreferenced items are excluded. This process occurs during VACUUM's second pass over the heap, right after LP_DEAD line pointers on the page (those encountered/pruned during the first pass) are marked LP_UNUSED. Truncation avoids line pointer bloat with certain workloads, particularly those involving continual range DELETEs and bulk INSERTs against the same table. Also harden heapam code to check for an out-of-range page offset number in places where we weren't already doing so. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAH2-Wzn6a64PJM1Ggzm=uvx2otsopJMhFQj_g1rAj4GWr3ZSzw@mail.gmail.com
2021-04-07 17:47:15 +02:00
* We may also be able to truncate the line pointer array of the heap pages we
* visit. If there is a contiguous group of LP_UNUSED items at the end of the
* array, it can be reclaimed as free space. These LP_UNUSED items usually
* start out as LP_DEAD items recorded by lazy_scan_prune (we set items from
* each page to LP_UNUSED, and then consider if it's possible to truncate the
* page's line pointer array).
*
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
* Note: the reason for doing this as a second pass is we cannot remove the
* tuples until we've removed their index entries, and we want to process
* index entry removal in batches as large as possible.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
static void
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_vacuum_heap_rel(LVRelState *vacrel)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Make lazy_vacuum_heap_rel match lazy_scan_heap. Make lazy_vacuum_heap_rel variable names match those from lazy_scan_heap where that makes sense. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 03:45:32 +01:00
int index = 0;
BlockNumber vacuumed_pages = 0;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
Buffer vmbuffer = InvalidBuffer;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
LVSavedErrInfo saved_err_info;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
Assert(vacrel->do_index_vacuuming);
Assert(vacrel->do_index_cleanup);
Assert(vacrel->num_index_scans > 0);
Refactor code dedicated to index vacuuming in vacuumlazy.c The part in charge of doing the vacuum on all the indexes of a relation was duplicated, with the same handling for progress reporting done. While on it, update the progress reporting for heap vacuuming in the subroutine doing the actual work, keeping the status update local. This way, any future caller of lazy_vacuum_heap() does not have to worry about doing any progress reporting update. Author: Justin Pryzby, Michael Paquier Discussion: https://postgr.es/m/20191120210600.GC30362@telsasoft.com
2019-12-26 09:01:23 +01:00
/* Report that we are now vacuuming the heap */
pgstat_progress_update_param(PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_PHASE_VACUUM_HEAP);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Update error traceback information */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
update_vacuum_error_info(vacrel, &saved_err_info,
VACUUM_ERRCB_PHASE_VACUUM_HEAP,
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
InvalidBlockNumber, InvalidOffsetNumber);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
while (index < vacrel->dead_items->num_items)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Make lazy_vacuum_heap_rel match lazy_scan_heap. Make lazy_vacuum_heap_rel variable names match those from lazy_scan_heap where that makes sense. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 03:45:32 +01:00
BlockNumber blkno;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Buffer buf;
Page page;
Rewrite the FSM. Instead of relying on a fixed-size shared memory segment, the free space information is stored in a dedicated FSM relation fork, with each relation (except for hash indexes; they don't use FSM). This eliminates the max_fsm_relations and max_fsm_pages GUC options; remove any trace of them from the backend, initdb, and documentation. Rewrite contrib/pg_freespacemap to match the new FSM implementation. Also introduce a new variant of the get_raw_page(regclass, int4, int4) function in contrib/pageinspect that let's you to return pages from any relation fork, and a new fsm_page_contents() function to inspect the new FSM pages.
2008-09-30 12:52:14 +02:00
Size freespace;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Centralize implementation of delay code by creating a pg_usleep() subroutine in src/port/pgsleep.c. Remove platform dependencies from miscadmin.h and put them in port.h where they belong. Extend recent vacuum cost-based-delay patch to apply to VACUUM FULL, ANALYZE, and non-btree index vacuuming. By the way, where is the documentation for the cost-based-delay patch?
2004-02-10 04:42:45 +01:00
vacuum_delay_point();
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
Make lazy_vacuum_heap_rel match lazy_scan_heap. Make lazy_vacuum_heap_rel variable names match those from lazy_scan_heap where that makes sense. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 03:45:32 +01:00
blkno = ItemPointerGetBlockNumber(&vacrel->dead_items->items[index]);
vacrel->blkno = blkno;
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
/*
* Pin the visibility map page in case we need to mark the page
* all-visible. In most cases this will be very cheap, because we'll
* already have the correct page pinned anyway.
*/
visibilitymap_pin(vacrel->rel, blkno, &vmbuffer);
/* We need a non-cleanup exclusive lock to mark dead_items unused */
Make lazy_vacuum_heap_rel match lazy_scan_heap. Make lazy_vacuum_heap_rel variable names match those from lazy_scan_heap where that makes sense. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 03:45:32 +01:00
buf = ReadBufferExtended(vacrel->rel, MAIN_FORKNUM, blkno, RBM_NORMAL,
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->bstrategy);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
index = lazy_vacuum_heap_page(vacrel, blkno, buf, index, vmbuffer);
Rewrite the FSM. Instead of relying on a fixed-size shared memory segment, the free space information is stored in a dedicated FSM relation fork, with each relation (except for hash indexes; they don't use FSM). This eliminates the max_fsm_relations and max_fsm_pages GUC options; remove any trace of them from the backend, initdb, and documentation. Rewrite contrib/pg_freespacemap to match the new FSM implementation. Also introduce a new variant of the get_raw_page(regclass, int4, int4) function in contrib/pageinspect that let's you to return pages from any relation fork, and a new fsm_page_contents() function to inspect the new FSM pages.
2008-09-30 12:52:14 +02:00
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
/* Now that we've vacuumed the page, record its available space */
Revert no-op changes to BufferGetPage() The reverted changes were intended to force a choice of whether any newly-added BufferGetPage() calls needed to be accompanied by a test of the snapshot age, to support the "snapshot too old" feature. Such an accompanying test is needed in about 7% of the cases, where the page is being used as part of a scan rather than positioning for other purposes (such as DML or vacuuming). The additional effort required for back-patching, and the doubt whether the intended benefit would really be there, have indicated it is best just to rely on developers to do the right thing based on comments and existing usage, as we do with many other conventions. This change should have little or no effect on generated executable code. Motivated by the back-patching pain of Tom Lane and Robert Haas
2016-04-20 15:31:19 +02:00
page = BufferGetPage(buf);
Rewrite the FSM. Instead of relying on a fixed-size shared memory segment, the free space information is stored in a dedicated FSM relation fork, with each relation (except for hash indexes; they don't use FSM). This eliminates the max_fsm_relations and max_fsm_pages GUC options; remove any trace of them from the backend, initdb, and documentation. Rewrite contrib/pg_freespacemap to match the new FSM implementation. Also introduce a new variant of the get_raw_page(regclass, int4, int4) function in contrib/pageinspect that let's you to return pages from any relation fork, and a new fsm_page_contents() function to inspect the new FSM pages.
2008-09-30 12:52:14 +02:00
freespace = PageGetHeapFreeSpace(page);
Clean up WAL/buffer interactions as per my recent proposal. Get rid of the misleadingly-named WriteBuffer routine, and instead require routines that change buffer pages to call MarkBufferDirty (which does exactly what it says). We also require that they do so before calling XLogInsert; this takes care of the synchronization requirement documented in SyncOneBuffer. Note that because bufmgr takes the buffer content lock (in shared mode) while writing out any buffer, it doesn't matter whether MarkBufferDirty is executed before the buffer content change is complete, so long as the content change is completed before releasing exclusive lock on the buffer. So it's OK to set the dirtybit before we fill in the LSN. This eliminates the former kluge of needing to set the dirtybit in LockBuffer. Aside from making the code more transparent, we can also add some new debugging assertions, in particular that the caller of MarkBufferDirty must hold the buffer content lock, not merely a pin.
2006-04-01 01:32:07 +02:00
UnlockReleaseBuffer(buf);
Make lazy_vacuum_heap_rel match lazy_scan_heap. Make lazy_vacuum_heap_rel variable names match those from lazy_scan_heap where that makes sense. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 03:45:32 +01:00
RecordPageWithFreeSpace(vacrel->rel, blkno, freespace);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacuumed_pages++;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->blkno = InvalidBlockNumber;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
if (BufferIsValid(vmbuffer))
ReleaseBuffer(vmbuffer);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
/*
* We set all LP_DEAD items from the first heap pass to LP_UNUSED during
* the second heap pass. No more, no less.
*/
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
Assert(index > 0);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
Assert(vacrel->num_index_scans > 1 ||
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
(index == vacrel->lpdead_items &&
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
vacuumed_pages == vacrel->lpdead_item_pages));
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ereport(DEBUG2,
Message style improvements
2021-08-07 12:09:22 +02:00
(errmsg("table \"%s\": removed %lld dead item identifiers in %u pages",
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
vacrel->relname, (long long) index, vacuumed_pages)));
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Revert to the previous phase information for error traceback */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
restore_vacuum_error_info(vacrel, &saved_err_info);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
/*
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
* lazy_vacuum_heap_page() -- free page's LP_DEAD items listed in the
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* vacrel->dead_items array.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
vacuumlazy.c: prefer the term "cleanup lock". The term "super-exclusive lock" is an acceptable synonym of "cleanup lock". Even still, switching from one term to the other in the same file is confusing. Standardize on "cleanup lock" within vacuumlazy.c. Per a complaint from Andres Freund.
2021-11-28 01:05:01 +01:00
* Caller must have an exclusive buffer lock on the buffer (though a full
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
* cleanup lock is also acceptable). vmbuffer must be valid and already have
* a pin on blkno's visibility map page.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* index is an offset into the vacrel->dead_items array for the first listed
* LP_DEAD item on the page. The return value is the first index immediately
* after all LP_DEAD items for the same page in the array.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
static int
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_vacuum_heap_page(LVRelState *vacrel, BlockNumber blkno, Buffer buffer,
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
int index, Buffer vmbuffer)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
VacDeadItems *dead_items = vacrel->dead_items;
Revert no-op changes to BufferGetPage() The reverted changes were intended to force a choice of whether any newly-added BufferGetPage() calls needed to be accompanied by a test of the snapshot age, to support the "snapshot too old" feature. Such an accompanying test is needed in about 7% of the cases, where the page is being used as part of a scan rather than positioning for other purposes (such as DML or vacuuming). The additional effort required for back-patching, and the doubt whether the intended benefit would really be there, have indicated it is best just to rely on developers to do the right thing based on comments and existing usage, as we do with many other conventions. This change should have little or no effect on generated executable code. Motivated by the back-patching pain of Tom Lane and Robert Haas
2016-04-20 15:31:19 +02:00
Page page = BufferGetPage(buffer);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
OffsetNumber unused[MaxHeapTuplesPerPage];
vacuumlazy.c: Tweak local variable name. Make a local variable name consistent with the name from its WAL record. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 02:57:18 +01:00
int nunused = 0;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
TransactionId visibility_cutoff_xid;
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
bool all_frozen;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
LVSavedErrInfo saved_err_info;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Refactor lazy_scan_heap() loop. Add a lazy_scan_heap() subsidiary function that handles heap pruning and tuple freezing: lazy_scan_prune(). This is a great deal cleaner. The code that remains in lazy_scan_heap()'s per-block loop can now be thought of as code that either comes before or after the call to lazy_scan_prune(), which is now the clear focal point. This division is enforced by the way in which we now manage state. lazy_scan_prune() outputs state (using its own struct) that describes what to do with the page following pruning and freezing (e.g., visibility map maintenance, recording free space in the FSM). It doesn't get passed any special instructional state from the preamble code, though. Also cleanly separate the logic used by a VACUUM with INDEX_CLEANUP=off from the logic used by single-heap-pass VACUUMs. The former case is now structured as the omission of index and heap vacuuming by a two pass VACUUM. The latter case goes back to being used only when the table happens to have no indexes (just as it was before commit a96c41fe). This structure is much more natural, since the whole point of INDEX_CLEANUP=off is to skip the index and heap vacuuming that would otherwise take place. The single-heap-pass case doesn't skip any useful work, though -- it just does heap pruning and heap vacuuming together when the table happens to have no indexes. Both of these changes are preparation for an upcoming patch that generalizes the mechanism used by INDEX_CLEANUP=off. The later patch will allow VACUUM to give up on index and heap vacuuming dynamically, as problems emerge (e.g., with wraparound), so that an affected VACUUM operation can finish up as soon as possible. Also fix a very old bug in single-pass VACUUM VERBOSE output. We were reporting the number of tuples deleted via pruning as a direct substitute for reporting the number of LP_DEAD items removed in a function that deals with the second pass over the heap. But that doesn't work at all -- they're two different things. To fix, start tracking the total number of LP_DEAD items encountered during pruning, and use that in the report instead. A single pass VACUUM will always vacuum away whatever LP_DEAD items a heap page has immediately after it is pruned, so the total number of LP_DEAD items encountered during pruning equals the total number vacuumed-away. (They are _not_ equal in the INDEX_CLEANUP=off case, but that's okay because skipping index vacuuming is now a totally orthogonal concept to one-pass VACUUM.) Also stop reporting the count of LP_UNUSED items in VACUUM VERBOSE output. This makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (because it never showed information about LP_UNUSED items). VACUUM VERBOSE reported LP_UNUSED items left behind by the last VACUUM, and LP_UNUSED items created via pruning HOT chains during the current VACUUM (it never included LP_UNUSED items left behind by the current VACUUM's second pass over the heap). This makes it useless as an indicator of line pointer bloat, which must have been the original intention. (Like the first VACUUM VERBOSE issue, this issue was arguably an oversight in commit 282d2a03, which added the heap-only tuple optimization.) Finally, stop reporting empty_pages in VACUUM VERBOSE output, and start reporting pages_removed instead. This also makes the output of VACUUM VERBOSE more consistent with log_autovacuum's output (which does not show empty_pages, but does show pages_removed). An empty page isn't meaningfully different to a page that is almost empty, or a page that is empty but for only a small number of remaining LP_UNUSED items. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 16:49:39 +02:00
Assert(vacrel->nindexes == 0 || vacrel->do_index_vacuuming);
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
pgstat_progress_update_param(PROGRESS_VACUUM_HEAP_BLKS_VACUUMED, blkno);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Update error traceback information */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
update_vacuum_error_info(vacrel, &saved_err_info,
VACUUM_ERRCB_PHASE_VACUUM_HEAP, blkno,
InvalidOffsetNumber);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
START_CRIT_SECTION();
Clean up WAL/buffer interactions as per my recent proposal. Get rid of the misleadingly-named WriteBuffer routine, and instead require routines that change buffer pages to call MarkBufferDirty (which does exactly what it says). We also require that they do so before calling XLogInsert; this takes care of the synchronization requirement documented in SyncOneBuffer. Note that because bufmgr takes the buffer content lock (in shared mode) while writing out any buffer, it doesn't matter whether MarkBufferDirty is executed before the buffer content change is complete, so long as the content change is completed before releasing exclusive lock on the buffer. So it's OK to set the dirtybit before we fill in the LSN. This eliminates the former kluge of needing to set the dirtybit in LockBuffer. Aside from making the code more transparent, we can also add some new debugging assertions, in particular that the caller of MarkBufferDirty must hold the buffer content lock, not merely a pin.
2006-04-01 01:32:07 +02:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
for (; index < dead_items->num_items; index++)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
BlockNumber tblk;
OffsetNumber toff;
HOT updates. When we update a tuple without changing any of its indexed columns, and the new version can be stored on the same heap page, we no longer generate extra index entries for the new version. Instead, index searches follow the HOT-chain links to ensure they find the correct tuple version. In addition, this patch introduces the ability to "prune" dead tuples on a per-page basis, without having to do a complete VACUUM pass to recover space. VACUUM is still needed to clean up dead index entries, however. Pavan Deolasee, with help from a bunch of other people.
2007-09-20 19:56:33 +02:00
ItemId itemid;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
tblk = ItemPointerGetBlockNumber(&dead_items->items[index]);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
if (tblk != blkno)
break; /* past end of tuples for this block */
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
toff = ItemPointerGetOffsetNumber(&dead_items->items[index]);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
itemid = PageGetItemId(page, toff);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
Assert(ItemIdIsDead(itemid) && !ItemIdHasStorage(itemid));
Redefine the lp_flags field of item pointers as having four states, rather than two independent bits (one of which was never used in heap pages anyway, or at least hadn't been in a very long time). This gives us flexibility to add the HOT notions of redirected and dead item pointers without requiring anything so klugy as magic values of lp_off and lp_len. The state values are chosen so that for the states currently in use (pre-HOT) there is no change in the physical representation.
2007-09-13 00:10:26 +02:00
ItemIdSetUnused(itemid);
vacuumlazy.c: Tweak local variable name. Make a local variable name consistent with the name from its WAL record. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 02:57:18 +01:00
unused[nunused++] = toff;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
vacuumlazy.c: Tweak local variable name. Make a local variable name consistent with the name from its WAL record. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 02:57:18 +01:00
Assert(nunused > 0);
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
Truncate line pointer array during VACUUM. Teach VACUUM to truncate the line pointer array of each heap page when a contiguous group of LP_UNUSED line pointers appear at the end of the array -- these unused and unreferenced items are excluded. This process occurs during VACUUM's second pass over the heap, right after LP_DEAD line pointers on the page (those encountered/pruned during the first pass) are marked LP_UNUSED. Truncation avoids line pointer bloat with certain workloads, particularly those involving continual range DELETEs and bulk INSERTs against the same table. Also harden heapam code to check for an out-of-range page offset number in places where we weren't already doing so. Author: Matthias van de Meent <boekewurm+postgres@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAEze2WjgaQc55Y5f5CQd3L=eS5CZcff2Obxp=O6pto8-f0hC4w@mail.gmail.com Discussion: https://postgr.es/m/CAH2-Wzn6a64PJM1Ggzm=uvx2otsopJMhFQj_g1rAj4GWr3ZSzw@mail.gmail.com
2021-04-07 17:47:15 +02:00
/* Attempt to truncate line pointer array now */
PageTruncateLinePointerArray(page);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Ensure we MarkBufferDirty before visibilitymap_set() logs the heap page and sets the LSN. Otherwise a checkpoint could occur between those actions and leave us in an inconsistent state. Jeff Davis
2013-04-30 09:15:49 +02:00
/*
* Mark buffer dirty before we write WAL.
*/
MarkBufferDirty(buffer);
Fix WAL-logging of setting the visibility map bit. The operation that removes the remaining dead tuples from the page must be WAL-logged before the setting of the VM bit. Otherwise, if you replay the WAL to between those two records, you end up with the VM bit set, but the dead tuples are still there. Backpatch to 9.3, where this bug was introduced.
2013-12-13 12:52:47 +01:00
/* XLOG stuff */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (RelationNeedsWAL(vacrel->rel))
Fix WAL-logging of setting the visibility map bit. The operation that removes the remaining dead tuples from the page must be WAL-logged before the setting of the VM bit. Otherwise, if you replay the WAL to between those two records, you end up with the VM bit set, but the dead tuples are still there. Backpatch to 9.3, where this bug was introduced.
2013-12-13 12:52:47 +01:00
{
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
xl_heap_vacuum xlrec;
Fix WAL-logging of setting the visibility map bit. The operation that removes the remaining dead tuples from the page must be WAL-logged before the setting of the VM bit. Otherwise, if you replay the WAL to between those two records, you end up with the VM bit set, but the dead tuples are still there. Backpatch to 9.3, where this bug was introduced.
2013-12-13 12:52:47 +01:00
XLogRecPtr recptr;
vacuumlazy.c: Tweak local variable name. Make a local variable name consistent with the name from its WAL record. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 02:57:18 +01:00
xlrec.nunused = nunused;
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
XLogBeginInsert();
XLogRegisterData((char *) &xlrec, SizeOfHeapVacuum);
XLogRegisterBuffer(0, buffer, REGBUF_STANDARD);
vacuumlazy.c: Tweak local variable name. Make a local variable name consistent with the name from its WAL record. Extracted from a larger patch to deal with issues with how vacuumlazy.c sets pages all-frozen. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-12 02:57:18 +01:00
XLogRegisterBufData(0, (char *) unused, nunused * sizeof(OffsetNumber));
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
recptr = XLogInsert(RM_HEAP2_ID, XLOG_HEAP2_VACUUM);
Fix WAL-logging of setting the visibility map bit. The operation that removes the remaining dead tuples from the page must be WAL-logged before the setting of the VM bit. Otherwise, if you replay the WAL to between those two records, you end up with the VM bit set, but the dead tuples are still there. Backpatch to 9.3, where this bug was introduced.
2013-12-13 12:52:47 +01:00
PageSetLSN(page, recptr);
}
Do all-visible handling in lazy_vacuum_page() outside its critical section. Since fdf9e21196a lazy_vacuum_page() rechecks the all-visible status of pages in the second pass over the heap. It does so inside a critical section, but both visibilitymap_test() and heap_page_is_all_visible() perform operations that should not happen inside one. The former potentially performs IO and both potentially do memory allocations. To fix, simply move all the all-visible handling outside the critical section. Doing so means that the PD_ALL_VISIBLE on the page won't be included in the full page image of the HEAP2_CLEAN record anymore. But that's fine, the flag will be set by the HEAP2_VISIBLE logged later. Backpatch to 9.3 where the problem was introduced. The bug only came to light due to the assertion added in 4a170ee9 and isn't likely to cause problems in production scenarios. The worst outcome is a avoidable PANIC restart. This also gets rid of the difference in the order of operations between master and standby mentioned in 2a8e1ac5. Per reports from David Leverton and Keith Fiske in bug #10533.
2014-06-20 11:06:42 +02:00
/*
* End critical section, so we safely can do visibility tests (which
* possibly need to perform IO and allocate memory!). If we crash now the
* page (including the corresponding vm bit) might not be marked all
* visible, but that's fine. A later vacuum will fix that.
*/
END_CRIT_SECTION();
/*
Fix typos in comments The changes done in this commit impact comments with no direct user-visible changes, with fixes for incorrect function, variable or structure names. Author: Alexander Lakhin Discussion: https://postgr.es/m/e8c38840-596a-83d6-bd8d-cebc51111572@gmail.com
2023-05-02 05:23:08 +02:00
* Now that we have removed the LP_DEAD items from the page, once again
Do all-visible handling in lazy_vacuum_page() outside its critical section. Since fdf9e21196a lazy_vacuum_page() rechecks the all-visible status of pages in the second pass over the heap. It does so inside a critical section, but both visibilitymap_test() and heap_page_is_all_visible() perform operations that should not happen inside one. The former potentially performs IO and both potentially do memory allocations. To fix, simply move all the all-visible handling outside the critical section. Doing so means that the PD_ALL_VISIBLE on the page won't be included in the full page image of the HEAP2_CLEAN record anymore. But that's fine, the flag will be set by the HEAP2_VISIBLE logged later. Backpatch to 9.3 where the problem was introduced. The bug only came to light due to the assertion added in 4a170ee9 and isn't likely to cause problems in production scenarios. The worst outcome is a avoidable PANIC restart. This also gets rid of the difference in the order of operations between master and standby mentioned in 2a8e1ac5. Per reports from David Leverton and Keith Fiske in bug #10533.
2014-06-20 11:06:42 +02:00
* check if the page has become all-visible. The page is already marked
* dirty, exclusively locked, and, if needed, a full page image has been
Remove tupgone special case from vacuumlazy.c. Retry the call to heap_prune_page() in rare cases where there is disagreement between the heap_prune_page() call and the call to HeapTupleSatisfiesVacuum() that immediately follows. Disagreement is possible when a concurrently-aborted transaction makes a tuple DEAD during the tiny window between each step. This was the only case where a tuple considered DEAD by VACUUM still had storage following pruning. VACUUM's definition of dead tuples is now uniformly simple and unambiguous: dead tuples from each page are always LP_DEAD line pointers that were encountered just after we performed pruning (and just before we considered freezing remaining items with tuple storage). Eliminating the tupgone=true special case enables INDEX_CLEANUP=off style skipping of index vacuuming that takes place based on flexible, dynamic criteria. The INDEX_CLEANUP=off case had to know about skipping indexes up-front before now, due to a subtle interaction with the special case (see commit dd695979) -- this was a special case unto itself. Now there are no special cases. And so now it won't matter when or how we decide to skip index vacuuming: it won't affect how pruning behaves, and it won't be affected by any of the implementation details of pruning or freezing. Also remove XLOG_HEAP2_CLEANUP_INFO records. These are no longer necessary because we now rely entirely on heap pruning taking care of recovery conflicts. There is no longer any need to generate recovery conflicts for DEAD tuples that pruning just missed. This also means that heap vacuuming now uses exactly the same strategy for recovery conflicts as index vacuuming always has: REDO routines never need to process a latestRemovedXid from the WAL record, since earlier REDO of the WAL record from pruning is sufficient in all cases. The generic XLOG_HEAP2_CLEAN record type is now split into two new record types to reflect this new division (these are called XLOG_HEAP2_PRUNE and XLOG_HEAP2_VACUUM). Also stop acquiring a super-exclusive lock for heap pages when they're vacuumed during VACUUM's second heap pass. A regular exclusive lock is enough. This is correct because heap page vacuuming is now strictly a matter of setting the LP_DEAD line pointers to LP_UNUSED. No other backend can have a pointer to a tuple located in a pinned buffer that can be invalidated by a concurrent heap page vacuum operation. Heap vacuuming can now be thought of as conceptually similar to index vacuuming and conceptually dissimilar to heap pruning. Heap pruning now has sole responsibility for anything involving the logical contents of the database (e.g., managing transaction status information, recovery conflicts, considering what to do with HOT chains). Index vacuuming and heap vacuuming are now only concerned with recycling garbage items from physical data structures that back the logical database. Bump XLOG_PAGE_MAGIC due to pruning and heap page vacuum WAL record changes. Credit for the idea of retrying pruning a page to avoid the tupgone case goes to Andres Freund. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WznneCXTzuFmcwx_EyRQgfsfJAAsu+CsqRFmFXCAar=nJw@mail.gmail.com
2021-04-06 17:49:22 +02:00
* emitted.
Do all-visible handling in lazy_vacuum_page() outside its critical section. Since fdf9e21196a lazy_vacuum_page() rechecks the all-visible status of pages in the second pass over the heap. It does so inside a critical section, but both visibilitymap_test() and heap_page_is_all_visible() perform operations that should not happen inside one. The former potentially performs IO and both potentially do memory allocations. To fix, simply move all the all-visible handling outside the critical section. Doing so means that the PD_ALL_VISIBLE on the page won't be included in the full page image of the HEAP2_CLEAN record anymore. But that's fine, the flag will be set by the HEAP2_VISIBLE logged later. Backpatch to 9.3 where the problem was introduced. The bug only came to light due to the assertion added in 4a170ee9 and isn't likely to cause problems in production scenarios. The worst outcome is a avoidable PANIC restart. This also gets rid of the difference in the order of operations between master and standby mentioned in 2a8e1ac5. Per reports from David Leverton and Keith Fiske in bug #10533.
2014-06-20 11:06:42 +02:00
*/
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
Assert(!PageIsAllVisible(page));
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (heap_page_is_all_visible(vacrel, buffer, &visibility_cutoff_xid,
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
&all_frozen))
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
{
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
uint8 flags = VISIBILITYMAP_ALL_VISIBLE;
if (all_frozen)
{
Assert(!TransactionIdIsValid(visibility_cutoff_xid));
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
flags |= VISIBILITYMAP_ALL_FROZEN;
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
}
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
Tighten up VACUUM's approach to setting VM bits. Tighten up the way that visibilitymap_set() is called: request that both the all-visible and all-frozen bits get set whenever the all-frozen bit is set, regardless of what we think we know about the present state of the all-visible bit. Also make sure that the page level PD_ALL_VISIBLE flag is set in the same code path. In practice there doesn't seem to be a concrete scenario in which the previous approach could lead to inconsistencies. It was almost possible in scenarios involving concurrent HOT updates from transactions that abort, but (unlike pruning) freezing can never remove XIDs > VACUUM's OldestXmin, even those from transactions that are known to have aborted. That was protective here. These issues have been around since commit a892234f83, which added the all-frozen bit to the VM fork. There is no known live bug here, so no backpatch. In passing, add some defensive assertions to catch the issue, and stop reading the existing state of the VM when setting the VM in VACUUM's final heap pass. We already know that affected pages must have had at least one LP_DEAD item before we set it LP_UNUSED, so there is no point in reading the VM when it is set like this. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-16 18:34:37 +01:00
PageSetAllVisible(page);
visibilitymap_set(vacrel->rel, blkno, buffer, InvalidXLogRecPtr,
vmbuffer, visibility_cutoff_xid, flags);
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
}
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Revert to the previous phase information for error traceback */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
restore_vacuum_error_info(vacrel, &saved_err_info);
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
return index;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
/*
* Trigger the failsafe to avoid wraparound failure when vacrel table has a
* relfrozenxid and/or relminmxid that is dangerously far in the past.
* Triggering the failsafe makes the ongoing VACUUM bypass any further index
Don't truncate heap when VACUUM's failsafe is in effect. It seems like a good idea to bypass heap truncation when the wraparound failsafe mechanism (which was added in commit 1e55e7d1) is in effect. Deliberately don't bypass heap truncation in the INDEX_CLEANUP=off case, even though it is similar to the failsafe case. There is already a separate reloption (and related VACUUM parameter) for that. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAD21AoDWRh6oTN5T8wa+cpZUVpHXET8BJ8Da7WHVHpwkPP6KLg@mail.gmail.com
2021-04-13 21:58:31 +02:00
* vacuuming and heap vacuuming. Truncating the heap is also bypassed.
*
* Any remaining work (work that VACUUM cannot just bypass) is typically sped
* up when the failsafe triggers. VACUUM stops applying any cost-based delay
* that it started out with.
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
*
* Returns true when failsafe has been triggered.
*/
static bool
lazy_check_wraparound_failsafe(LVRelState *vacrel)
{
/* Don't warn more than once per VACUUM */
Make vacuum failsafe_active globally visible While vacuuming a table in failsafe mode, VacuumCostActive should not be re-enabled. This currently isn't a problem because vacuum cost parameters are only refreshed in between vacuuming tables and failsafe status is reset for every table. In preparation for allowing vacuum cost parameters to be updated more frequently, elevate LVRelState->failsafe_active to a global, VacuumFailsafeActive, which will be checked when determining whether or not to re-enable vacuum cost-related delays. Author: Melanie Plageman <melanieplageman@gmail.com> Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Daniel Gustafsson <daniel@yesql.se> Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Reviewed-by: Robert Haas <robertmhaas@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/CAAKRu_ZngzqnEODc7LmS1NH04Kt6Y9huSjz5pp7%2BDXhrjDA0gw%40mail.gmail.com
2023-04-07 00:54:08 +02:00
if (VacuumFailsafeActive)
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
return true;
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
if (unlikely(vacuum_xid_failsafe_check(&vacrel->cutoffs)))
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
{
Make vacuum failsafe_active globally visible While vacuuming a table in failsafe mode, VacuumCostActive should not be re-enabled. This currently isn't a problem because vacuum cost parameters are only refreshed in between vacuuming tables and failsafe status is reset for every table. In preparation for allowing vacuum cost parameters to be updated more frequently, elevate LVRelState->failsafe_active to a global, VacuumFailsafeActive, which will be checked when determining whether or not to re-enable vacuum cost-related delays. Author: Melanie Plageman <melanieplageman@gmail.com> Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Daniel Gustafsson <daniel@yesql.se> Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Reviewed-by: Robert Haas <robertmhaas@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/CAAKRu_ZngzqnEODc7LmS1NH04Kt6Y9huSjz5pp7%2BDXhrjDA0gw%40mail.gmail.com
2023-04-07 00:54:08 +02:00
VacuumFailsafeActive = true;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
Disable vacuum's use of a buffer access strategy during failsafe Traditionally, vacuum always makes use of a buffer access strategy 32 buffers in size. This means that running vacuums tend not to cause too many shared buffers to become dirty, however, this can cause vacuums to run much more slowly than they otherwise could as WAL flushes will occur more frequently due to having to flush WAL out to the LSN of the dirty page before that page can be written to disk. When we are performing failsafe VACUUMs (as added in 1e55e7d17), we really want to make the vacuum work go as quickly as possible, so here we disable the buffer access strategy when entering failsafe mode while vacuuming a relation. Per idea and analyis from Andres Freund. In passing, also include some changes I had intended for 32fbe0239. Author: Melanie Plageman Reviewed-by: Justin Pryzby, David Rowley Discussion: https://postgr.es/m/20230111182720.ejifsclfwymw2reb%40awork3.anarazel.de
2023-04-03 13:05:58 +02:00
/*
* Abandon use of a buffer access strategy to allow use of all of
* shared buffers. We assume the caller who allocated the memory for
* the BufferAccessStrategy will free it.
*/
vacrel->bstrategy = NULL;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
/* Disable index vacuuming, index cleanup, and heap rel truncation */
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
vacrel->do_index_vacuuming = false;
vacrel->do_index_cleanup = false;
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
vacrel->do_rel_truncate = false;
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
ereport(WARNING,
Consider triggering VACUUM failsafe during scan. The wraparound failsafe mechanism added by commit 1e55e7d1 handled the one-pass strategy case (i.e. the "table has no indexes" case) by adding a dedicated failsafe check. This made up for the fact that the usual one-pass checks inside lazy_vacuum_all_indexes() cannot ever be reached during a one-pass strategy VACUUM. This approach failed to account for two-pass VACUUMs that opt out of index vacuuming up-front. The INDEX_CLEANUP off case in the only case that works like that. Fix this by performing a failsafe check every 4GB during the first scan of the heap, regardless of the details of the VACUUM. This eliminates the special case, and will make the failsafe trigger more reliably. Author: Peter Geoghegan <pg@bowt.ie> Reported-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/20210424002921.pb3t7h6frupdqnkp@alap3.anarazel.de
2021-05-25 02:14:02 +02:00
(errmsg("bypassing nonessential maintenance of table \"%s.%s.%s\" as a failsafe after %d index scans",
vacuumlazy.c: Save get_database_name() in vacrel. This brings dbname strings in line with namespace and relation name strings. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WzkQ1TKU-DdNvnGeL870di3+CU1UTo-7nw7xFDpVE-XGjA@mail.gmail.com
2023-01-03 20:48:47 +01:00
vacrel->dbname, vacrel->relnamespace, vacrel->relname,
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
vacrel->num_index_scans),
Message style improvements
2021-08-07 12:09:22 +02:00
errdetail("The table's relfrozenxid or relminmxid is too far in the past."),
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
errhint("Consider increasing configuration parameter \"maintenance_work_mem\" or \"autovacuum_work_mem\".\n"
"You might also need to consider other ways for VACUUM to keep up with the allocation of transaction IDs.")));
/* Stop applying cost limits from this point on */
VacuumCostActive = false;
VacuumCostBalance = 0;
return true;
}
return false;
}
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
/*
* lazy_cleanup_all_indexes() -- cleanup all indexes of relation.
*/
static void
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_cleanup_all_indexes(LVRelState *vacrel)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
{
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
double reltuples = vacrel->new_rel_tuples;
bool estimated_count = vacrel->scanned_pages < vacrel->rel_pages;
Assert(vacrel->do_index_cleanup);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Assert(vacrel->nindexes > 0);
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
/* Report that we are now cleaning up indexes */
pgstat_progress_update_param(PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_PHASE_INDEX_CLEANUP);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (!ParallelVacuumIsActive(vacrel))
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
for (int idx = 0; idx < vacrel->nindexes; idx++)
{
Relation indrel = vacrel->indrels[idx];
IndexBulkDeleteResult *istat = vacrel->indstats[idx];
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->indstats[idx] =
lazy_cleanup_one_index(indrel, istat, reltuples,
estimated_count, vacrel);
}
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
}
else
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/* Outsource everything to parallel variant */
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
parallel_vacuum_cleanup_all_indexes(vacrel->pvs, reltuples,
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
vacrel->num_index_scans,
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
estimated_count);
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
}
}
Make VACUUM avoid waiting for a cleanup lock, where possible. In a regular VACUUM, it's OK to skip pages for which a cleanup lock isn't immediately available; the next VACUUM will deal with them. If we're scanning the entire relation to advance relfrozenxid, we might need to wait, but only if there are tuples on the page that actually require freezing. These changes should greatly reduce the incidence of of vacuum processes getting "stuck". Simon Riggs and Robert Haas
2011-11-08 03:39:40 +01:00
New-style vacuum neglected to update pg_class statistics about indexes if there were no deletions to do.
2001-07-18 02:46:25 +02:00
/*
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
* lazy_vacuum_one_index() -- vacuum index relation.
New-style vacuum neglected to update pg_class statistics about indexes if there were no deletions to do.
2001-07-18 02:46:25 +02:00
*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* Delete all the index tuples containing a TID collected in
* vacrel->dead_items array. Also update running statistics.
* Exact details depend on index AM's ambulkdelete routine.
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
*
* reltuples is the number of heap tuples to be passed to the
Redefine pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Historically, we've considered the state with relpages and reltuples both zero as indicating that we do not know the table's tuple density. This is problematic because it's impossible to distinguish "never yet vacuumed" from "vacuumed and seen to be empty". In particular, a user cannot use VACUUM or ANALYZE to override the planner's normal heuristic that an empty table should not be believed to be empty because it is probably about to get populated. That heuristic is a good safety measure, so I don't care to abandon it, but there should be a way to override it if the table is indeed intended to stay empty. Hence, represent the initial state of ignorance by setting reltuples to -1 (relpages is still set to zero), and apply the minimum-ten-pages heuristic only when reltuples is still -1. If the table is empty, VACUUM or ANALYZE (but not CREATE INDEX) will override that to reltuples = relpages = 0, and then we'll plan on that basis. This requires a bunch of fiddly little changes, but we can get rid of some ugly kluges that were formerly needed to maintain the old definition. One notable point is that FDWs' GetForeignRelSize methods will see baserel->tuples = -1 when no ANALYZE has been done on the foreign table. That seems like a net improvement, since those methods were formerly also in the dark about what baserel->tuples = 0 really meant. Still, it is an API change. I bumped catversion because code predating this change would get confused by seeing reltuples = -1. Discussion: https://postgr.es/m/F02298E0-6EF4-49A1-BCB6-C484794D9ACC@thebuild.com
2020-08-30 18:21:51 +02:00
* bulkdelete callback. It's always assumed to be estimated.
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* See indexam.sgml for more info.
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
*
* Returns bulk delete stats derived from input stats
New-style vacuum neglected to update pg_class statistics about indexes if there were no deletions to do.
2001-07-18 02:46:25 +02:00
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static IndexBulkDeleteResult *
lazy_vacuum_one_index(Relation indrel, IndexBulkDeleteResult *istat,
double reltuples, LVRelState *vacrel)
New-style vacuum neglected to update pg_class statistics about indexes if there were no deletions to do.
2001-07-18 02:46:25 +02:00
{
Clean up API for ambulkdelete/amvacuumcleanup as per today's discussion. This formulation requires every AM to provide amvacuumcleanup, unlike before, but it's surely a whole lot cleaner. Also, add an 'amstorage' column to pg_am so that we can get rid of hardwired knowledge in DefineOpClass().
2006-05-03 00:25:10 +02:00
IndexVacuumInfo ivinfo;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
LVSavedErrInfo saved_err_info;
New-style vacuum neglected to update pg_class statistics about indexes if there were no deletions to do.
2001-07-18 02:46:25 +02:00
Clean up API for ambulkdelete/amvacuumcleanup as per today's discussion. This formulation requires every AM to provide amvacuumcleanup, unlike before, but it's surely a whole lot cleaner. Also, add an 'amstorage' column to pg_am so that we can get rid of hardwired knowledge in DefineOpClass().
2006-05-03 00:25:10 +02:00
ivinfo.index = indrel;
Move heaprel struct field next to index rel field. Commit 61b313e4 added a heaprel struct member to IndexVacuumInfo, but placed it last. Move the heaprel struct member next to the index struct member to improve the code's readability. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznG=TV6S9d3VA=y0vBHbXwnLs9_LLdiML=aNJuHeriwxg@mail.gmail.com
2023-04-03 20:01:11 +02:00
ivinfo.heaprel = vacrel->rel;
Implement "fastupdate" support for GIN indexes, in which we try to accumulate multiple index entries in a holding area before adding them to the main index structure. This helps because bulk insert is (usually) significantly faster than retail insert for GIN. This patch also removes GIN support for amgettuple-style index scans. The API defined for amgettuple is difficult to support with fastupdate, and the previously committed partial-match feature didn't really work with it either. We might eventually figure a way to put back amgettuple support, but it won't happen for 8.4. catversion bumped because of change in GIN's pg_am entry, and because the format of GIN indexes changed on-disk (there's a metapage now, and possibly a pending list). Teodor Sigaev
2009-03-24 21:17:18 +01:00
ivinfo.analyze_only = false;
Correctly initialize newly added struct member Valgrind was rightly complaining that IndexVacuumInfo->report_progress (added by commit ab0dfc961b6a) was not being initialized in some code paths. Repair. Per buildfarm member lousyjack.
2019-04-03 14:56:20 +02:00
ivinfo.report_progress = false;
Improve the IndexVacuumInfo/IndexBulkDeleteResult API to allow somewhat sane behavior in cases where we don't know the heap tuple count accurately; in particular partial vacuum, but this also makes the API a bit more useful for ANALYZE. This patch adds "estimated_count" flags to both structs so that an approximate count can be flagged as such, and adjusts the logic so that approximate counts are not used for updating pg_class.reltuples. This fixes my previous complaint that VACUUM was putting ridiculous values into pg_class.reltuples for indexes. The actual impact of that bug is limited, because the planner only pays attention to reltuples for an index if the index is partial; which probably explains why beta testers hadn't noticed a degradation in plan quality from it. But it needs to be fixed. The whole thing is a bit messy and should be redesigned in future, because reltuples now has the potential to drift quite far away from reality when a long period elapses with no non-partial vacuums. But this is as good as it's going to get for 8.4.
2009-06-07 00:13:52 +02:00
ivinfo.estimated_count = true;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ivinfo.message_level = DEBUG2;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
ivinfo.num_heap_tuples = reltuples;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
ivinfo.strategy = vacrel->bstrategy;
More infrastructure for btree compaction project. Tree-traversal code now knows what to do upon hitting a dead page (in theory anyway, it's untested...). Add a post-VACUUM-cleanup entry point for index AMs, to provide a place for dead-page scavenging to happen. Also, fix oversight that broke btpo_prev links in temporary indexes. initdb forced due to additions in pg_am.
2003-02-22 01:45:05 +01:00
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
/*
* Update error traceback information.
*
* The index name is saved during this phase and restored immediately
* after this phase. See vacuum_error_callback.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Assert(vacrel->indname == NULL);
vacrel->indname = pstrdup(RelationGetRelationName(indrel));
update_vacuum_error_info(vacrel, &saved_err_info,
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
VACUUM_ERRCB_PHASE_VACUUM_INDEX,
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
InvalidBlockNumber, InvalidOffsetNumber);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
Clean up API for ambulkdelete/amvacuumcleanup as per today's discussion. This formulation requires every AM to provide amvacuumcleanup, unlike before, but it's surely a whole lot cleaner. Also, add an 'amstorage' column to pg_am so that we can get rid of hardwired knowledge in DefineOpClass().
2006-05-03 00:25:10 +02:00
/* Do bulk deletion */
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
istat = vac_bulkdel_one_index(&ivinfo, istat, (void *) vacrel->dead_items);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
/* Revert to the previous phase information for error traceback */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
restore_vacuum_error_info(vacrel, &saved_err_info);
pfree(vacrel->indname);
vacrel->indname = NULL;
return istat;
New-style vacuum neglected to update pg_class statistics about indexes if there were no deletions to do.
2001-07-18 02:46:25 +02:00
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
* lazy_cleanup_one_index() -- do post-vacuum cleanup for index relation.
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* Calls index AM's amvacuumcleanup routine. reltuples is the number
* of heap tuples and estimated_count is true if reltuples is an
* estimated value. See indexam.sgml for more info.
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
*
* Returns bulk delete stats derived from input stats
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
static IndexBulkDeleteResult *
lazy_cleanup_one_index(Relation indrel, IndexBulkDeleteResult *istat,
double reltuples, bool estimated_count,
LVRelState *vacrel)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Clean up API for ambulkdelete/amvacuumcleanup as per today's discussion. This formulation requires every AM to provide amvacuumcleanup, unlike before, but it's surely a whole lot cleaner. Also, add an 'amstorage' column to pg_am so that we can get rid of hardwired knowledge in DefineOpClass().
2006-05-03 00:25:10 +02:00
IndexVacuumInfo ivinfo;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
LVSavedErrInfo saved_err_info;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Clean up API for ambulkdelete/amvacuumcleanup as per today's discussion. This formulation requires every AM to provide amvacuumcleanup, unlike before, but it's surely a whole lot cleaner. Also, add an 'amstorage' column to pg_am so that we can get rid of hardwired knowledge in DefineOpClass().
2006-05-03 00:25:10 +02:00
ivinfo.index = indrel;
Move heaprel struct field next to index rel field. Commit 61b313e4 added a heaprel struct member to IndexVacuumInfo, but placed it last. Move the heaprel struct member next to the index struct member to improve the code's readability. Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznG=TV6S9d3VA=y0vBHbXwnLs9_LLdiML=aNJuHeriwxg@mail.gmail.com
2023-04-03 20:01:11 +02:00
ivinfo.heaprel = vacrel->rel;
Implement "fastupdate" support for GIN indexes, in which we try to accumulate multiple index entries in a holding area before adding them to the main index structure. This helps because bulk insert is (usually) significantly faster than retail insert for GIN. This patch also removes GIN support for amgettuple-style index scans. The API defined for amgettuple is difficult to support with fastupdate, and the previously committed partial-match feature didn't really work with it either. We might eventually figure a way to put back amgettuple support, but it won't happen for 8.4. catversion bumped because of change in GIN's pg_am entry, and because the format of GIN indexes changed on-disk (there's a metapage now, and possibly a pending list). Teodor Sigaev
2009-03-24 21:17:18 +01:00
ivinfo.analyze_only = false;
Correctly initialize newly added struct member Valgrind was rightly complaining that IndexVacuumInfo->report_progress (added by commit ab0dfc961b6a) was not being initialized in some code paths. Repair. Per buildfarm member lousyjack.
2019-04-03 14:56:20 +02:00
ivinfo.report_progress = false;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
ivinfo.estimated_count = estimated_count;
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ivinfo.message_level = DEBUG2;
Sync up our various ways of estimating pg_class.reltuples. VACUUM thought that reltuples represents the total number of tuples in the relation, while ANALYZE counted only live tuples. This can cause "flapping" in the value when background vacuums and analyzes happen separately. The planner's use of reltuples essentially assumes that it's the count of live (visible) tuples, so let's standardize on having it mean live tuples. Another issue is that the definition of "live tuple" isn't totally clear; what should be done with INSERT_IN_PROGRESS or DELETE_IN_PROGRESS tuples? ANALYZE's choices in this regard are made on the assumption that if the originating transaction commits at all, it will happen after ANALYZE finishes, so we should ignore the effects of the in-progress transaction --- unless it is our own transaction, and then we should count it. Let's propagate this definition into VACUUM, too. Likewise propagate this definition into CREATE INDEX, and into contrib/pgstattuple's pgstattuple_approx() function. Tomas Vondra, reviewed by Haribabu Kommi, some corrections by me Discussion: https://postgr.es/m/16db4468-edfa-830a-f921-39a50498e77e@2ndquadrant.com
2018-03-22 20:47:29 +01:00
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
ivinfo.num_heap_tuples = reltuples;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
ivinfo.strategy = vacrel->bstrategy;
More infrastructure for btree compaction project. Tree-traversal code now knows what to do upon hitting a dead page (in theory anyway, it's untested...). Add a post-VACUUM-cleanup entry point for index AMs, to provide a place for dead-page scavenging to happen. Also, fix oversight that broke btpo_prev links in temporary indexes. initdb forced due to additions in pg_am.
2003-02-22 01:45:05 +01:00
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
/*
* Update error traceback information.
*
* The index name is saved during this phase and restored immediately
* after this phase. See vacuum_error_callback.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Assert(vacrel->indname == NULL);
vacrel->indname = pstrdup(RelationGetRelationName(indrel));
update_vacuum_error_info(vacrel, &saved_err_info,
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
VACUUM_ERRCB_PHASE_INDEX_CLEANUP,
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
InvalidBlockNumber, InvalidOffsetNumber);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
istat = vac_cleanup_one_index(&ivinfo, istat);
Improve the vacuum error context phase information. We were displaying the wrong phase information for 'info' message in the index clean up phase because we were switching to the previous phase a bit early. We were also not displaying context information for heap phase unless the block number is valid which is fine for error cases but for messages at 'info' or lower error level it appears to be inconsistent with index phase information. Reported-by: Sawada Masahiko Author: Sawada Masahiko Reviewed-by: Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CA+fd4k4HcbhPnCs7paRTw1K-AHin8y4xKomB9Ru0ATw0UeTy2w@mail.gmail.com
2020-08-24 04:46:19 +02:00
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
/* Revert to the previous phase information for error traceback */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
restore_vacuum_error_info(vacrel, &saved_err_info);
pfree(vacrel->indname);
vacrel->indname = NULL;
return istat;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
/*
* should_attempt_truncation - should we attempt to truncate the heap?
*
* Don't even think about it unless we have a shot at releasing a goodly
Simplify lazy_scan_heap's handling of scanned pages. Redefine a scanned page as any heap page that actually gets pinned by VACUUM's first pass over the heap, regardless of whether or not the page was cleanup locked. Although it's fundamentally impossible to prune a heap page without a cleanup lock (since we cannot safely defragment the page), we can do just about everything else. The only notable further exception is freezing tuples, though even that is arguably a consequence of not being able to prune (not a separate issue). VACUUM now does as much of the same processing as possible for pages that could not be cleanup locked. Any failure to do specific required processing is treated as a special case exception, which will be rare in practice. We now collect any preexisting LP_DEAD items (left behind by earlier opportunistic pruning) in the dead_items array for these heap pages, and count their tuples in the usual way. Steps used to decide if we'll attempt relation truncation are performed in the usual way for no-cleanup-lock scanned pages, too. Although eliminating these special cases is intrinsically useful, it's even more useful as an enabler of further simplifications. The only essential difference between aggressive and non-aggressive is that only aggressive is _guaranteed_ to be able to advance relfrozenxid up to FreezeLimit. Advancing relfrozenxid is always useful, but before now non-aggressive VACUUMs threw away the opportunity to do so whenever a cleanup lock could not be acquired on any page, no matter what the details were. This was very pessimistic. It isn't actually necessary to "behave aggressively" to maintain the ability to advance relfrozenxid when a cleanup lock isn't immediately available (most of the time). The non-aggressive case will now make sure that it isn't safe to advance relfrozenxid (without waiting) using only a share lock. It will usually notice that there are no tuples that need to be frozen anyway, just like in the aggressive case -- and so it no longer wastes an opportunity to advance relfrozenxid over nothing. (The non-aggressive case still won't wait for a cleanup lock when there really are tuples on the page that need to be frozen, since that really would amount to "behaving aggressively".) VACUUM currently has a tendency to set heap pages to all-visible in the visibility map before it freezes all of the tuples on the page. Only a subsequent aggressive VACUUM will visit these pages to freeze their tuples, usually only when the tuple XIDs are much older than the vacuum_freeze_min_age GUC (FreezeLimit cutoff) is supposed to allow. And so non-aggressive VACUUMs are still far less likely to be able to advance relfrozenxid in practice, even with the enhancements from this commit. This remaining issue will be addressed by future work that overhauls the criteria for freezing tuples. Once that's in place, almost every VACUUM operation will be able to advance relfrozenxid in practice. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-11 23:32:17 +01:00
* number of pages. Otherwise, the time taken isn't worth it, mainly because
* an AccessExclusive lock must be replayed on any hot standby, where it can
* be particularly disruptive.
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
*
Update "don't truncate with failsafe" rationale. There is a very good (though non-obvious) reason to avoid relation truncation during a VACUUM that has triggered the failsafe mechanism, which was missed before now. Update related comments, so this isn't forgotten. Reported-By: John Naylor <john.naylor@enterprisedb.com> Discussion: https://postgr.es/m/CAFBsxsFiMPxQ-dHZ8tOgktn=+ffeJT3+GinZ4zdOGbmAnCYadA@mail.gmail.com
2022-02-16 00:16:19 +01:00
* Also don't attempt it if wraparound failsafe is in effect. The entire
* system might be refusing to allocate new XIDs at this point. The system
* definitely won't return to normal unless and until VACUUM actually advances
* the oldest relfrozenxid -- which hasn't happened for target rel just yet.
* If lazy_truncate_heap attempted to acquire an AccessExclusiveLock to
* truncate the table under these circumstances, an XID exhaustion error might
* make it impossible for VACUUM to fix the underlying XID exhaustion problem.
Document LP_DEAD accounting issues in VACUUM. Document VACUUM's soft assumption that any LP_DEAD items encountered during pruning will become LP_UNUSED items before VACUUM finishes up. This is integral to the accounting used by VACUUM to generate its final report on the table to the stats collector. It also affects how VACUUM determines which heap pages are truncatable. In both cases VACUUM is concerned with the likely contents of the page in the near future, not the current contents of the page. This state of affairs created the false impression that VACUUM's dead tuple accounting had significant difference with similar accounting used during ANALYZE. There were and are no substantive differences, at least when the soft assumption completely works out. This is far clearer now. Also document cases where things don't quite work out for VACUUM's dead tuple accounting. It's possible that a significant number of LP_DEAD items will be left behind by VACUUM, and won't be recorded as remaining dead tuples in VACUUM's statistics collector report. This behavior dates back to commit a96c41fe, which taught VACUUM to run without index and heap vacuuming at the user's request. The failsafe mechanism added to VACUUM more recently by commit 1e55e7d1 takes the same approach to dead tuple accounting. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-Wz=Jmtu18PrsYq3EvvZJGOmZqSO2u3bvKpx9xJa5uhNp=Q@mail.gmail.com
2021-04-20 03:55:31 +02:00
* There is very little chance of truncation working out when the failsafe is
* in effect in any case. lazy_scan_prune makes the optimistic assumption
* that any LP_DEAD items it encounters will always be LP_UNUSED by the time
* we're called.
Don't truncate heap when VACUUM's failsafe is in effect. It seems like a good idea to bypass heap truncation when the wraparound failsafe mechanism (which was added in commit 1e55e7d1) is in effect. Deliberately don't bypass heap truncation in the INDEX_CLEANUP=off case, even though it is similar to the failsafe case. There is already a separate reloption (and related VACUUM parameter) for that. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAD21AoDWRh6oTN5T8wa+cpZUVpHXET8BJ8Da7WHVHpwkPP6KLg@mail.gmail.com
2021-04-13 21:58:31 +02:00
*
Add comment & docs about no vacuum truncation with sto. Omission noted by Andres Freund.
2016-07-19 23:25:53 +02:00
* Also don't attempt it if we are doing early pruning/vacuuming, because a
* scan which cannot find a truncated heap page cannot determine that the
Don't truncate heap when VACUUM's failsafe is in effect. It seems like a good idea to bypass heap truncation when the wraparound failsafe mechanism (which was added in commit 1e55e7d1) is in effect. Deliberately don't bypass heap truncation in the INDEX_CLEANUP=off case, even though it is similar to the failsafe case. There is already a separate reloption (and related VACUUM parameter) for that. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAD21AoDWRh6oTN5T8wa+cpZUVpHXET8BJ8Da7WHVHpwkPP6KLg@mail.gmail.com
2021-04-13 21:58:31 +02:00
* snapshot is too old to read that page.
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
*/
static bool
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
should_attempt_truncation(LVRelState *vacrel)
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
{
BlockNumber possibly_freeable;
Make vacuum failsafe_active globally visible While vacuuming a table in failsafe mode, VacuumCostActive should not be re-enabled. This currently isn't a problem because vacuum cost parameters are only refreshed in between vacuuming tables and failsafe status is reset for every table. In preparation for allowing vacuum cost parameters to be updated more frequently, elevate LVRelState->failsafe_active to a global, VacuumFailsafeActive, which will be checked when determining whether or not to re-enable vacuum cost-related delays. Author: Melanie Plageman <melanieplageman@gmail.com> Reviewed-by: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-by: Daniel Gustafsson <daniel@yesql.se> Reviewed-by: Kyotaro Horiguchi <horikyota.ntt@gmail.com> Reviewed-by: Robert Haas <robertmhaas@gmail.com> Discussion: https://www.postgresql.org/message-id/flat/CAAKRu_ZngzqnEODc7LmS1NH04Kt6Y9huSjz5pp7%2BDXhrjDA0gw%40mail.gmail.com
2023-04-07 00:54:08 +02:00
if (!vacrel->do_rel_truncate || VacuumFailsafeActive ||
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
old_snapshot_threshold >= 0)
Don't truncate heap when VACUUM's failsafe is in effect. It seems like a good idea to bypass heap truncation when the wraparound failsafe mechanism (which was added in commit 1e55e7d1) is in effect. Deliberately don't bypass heap truncation in the INDEX_CLEANUP=off case, even though it is similar to the failsafe case. There is already a separate reloption (and related VACUUM parameter) for that. Reported-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAD21AoDWRh6oTN5T8wa+cpZUVpHXET8BJ8Da7WHVHpwkPP6KLg@mail.gmail.com
2021-04-13 21:58:31 +02:00
return false;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
possibly_freeable = vacrel->rel_pages - vacrel->nonempty_pages;
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
if (possibly_freeable > 0 &&
(possibly_freeable >= REL_TRUNCATE_MINIMUM ||
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
possibly_freeable >= vacrel->rel_pages / REL_TRUNCATE_FRACTION))
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
return true;
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
return false;
Avoid useless truncation attempts during VACUUM. VACUUM can skip heap pages altogether when there's a run of consecutive pages that are all-visible according to the visibility map. This causes it to not update its nonempty_pages count, just as if those pages were empty, which means that at the end we will think they are candidates for deletion. Thus, we may take the table's AccessExclusive lock only to find that no pages are really truncatable. This usually causes no real problems on a master server, thanks to the lock being acquired only conditionally; but on hot-standby servers, the same lock must be acquired unconditionally which can result in unnecessary query cancellations. To improve matters, force examination of the table's last page whenever we reach there with a nonempty_pages count that would allow a truncation attempt. If it's not empty, we'll advance nonempty_pages and thereby prevent the truncation attempt. If we are unable to acquire cleanup lock on that page, there's no need to force it, unless we're doing an anti-wraparound vacuum. We can just check for tuples with a shared buffer lock and then give up. (When we are doing an anti-wraparound vacuum, and decide it's okay to skip the page because it contains no freezable tuples, this patch still improves matters because nonempty_pages is properly updated, which it was not before.) Since only the last page is special-cased in this way, we might attempt a truncation that will release many fewer pages than the normal heuristic would suggest; at worst, only one page would be truncated. But that seems all right, because the situation won't repeat during the next vacuum. The real problem with the old logic is that the useless truncation attempt happens every time we vacuum, so long as the state of the last few dozen pages doesn't change. This is a longstanding deficiency, but since the consequences aren't very severe in most scenarios, I'm not going to risk a back-patch. Jeff Janes and Tom Lane
2015-12-30 23:13:15 +01:00
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
* lazy_truncate_heap - try to truncate off any empty pages at the end
*/
Fix up rickety handling of relation-truncation interlocks. Move rd_targblock, rd_fsm_nblocks, and rd_vm_nblocks from relcache to the smgr relation entries, so that they will get reset to InvalidBlockNumber whenever an smgr-level flush happens. Because we now send smgr invalidation messages immediately (not at end of transaction) when a relation truncation occurs, this ensures that other backends will reset their values before they next access the relation. We no longer need the unreliable assumption that a VACUUM that's doing a truncation will hold its AccessExclusive lock until commit --- in fact, we can intentionally release that lock as soon as we've completed the truncation. This patch therefore reverts (most of) Alvaro's patch of 2009-11-10, as well as my marginal hacking on it yesterday. We can also get rid of assorted no-longer-needed relcache flushes, which are far more expensive than an smgr flush because they kill a lot more state. In passing this patch fixes smgr_redo's failure to perform visibility-map truncation, and cleans up some rather dubious assumptions in freespace.c and visibilitymap.c about when rd_fsm_nblocks and rd_vm_nblocks can be out of date.
2010-02-09 22:43:30 +01:00
static void
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
lazy_truncate_heap(LVRelState *vacrel)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Remove unneeded old_rel_pages VACUUM state field. The field hasn't been used since commit 3d351d91, which redefined pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Also rename a local variable of the same name ("old_rel_pages"). This is used by relation truncation to represent the original relation size at the start of the ongoing VACUUM operation. Rename it to orig_rel_pages, since that's a lot clearer. (This name matches similar nearby code.)
2021-08-31 23:59:52 +02:00
BlockNumber orig_rel_pages = vacrel->rel_pages;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
BlockNumber new_rel_pages;
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
bool lock_waiter_detected;
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
int lock_retry;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Add simple VACUUM progress reporting. There's a lot more that could be done here yet - in particular, this reports only very coarse-grained information about the index vacuuming phase - but even as it stands, the new pg_stat_progress_vacuum can tell you quite a bit about what a long-running vacuum is actually doing. Amit Langote and Robert Haas, based on earlier work by Vinayak Pokale and Rahila Syed.
2016-03-15 18:31:18 +01:00
/* Report that we are now truncating */
pgstat_progress_update_param(PROGRESS_VACUUM_PHASE,
PROGRESS_VACUUM_PHASE_TRUNCATE);
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
/* Update error traceback information one last time */
update_vacuum_error_info(vacrel, NULL, VACUUM_ERRCB_PHASE_TRUNCATE,
vacrel->nonempty_pages, InvalidOffsetNumber);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/*
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
* Loop until no more truncating can be done.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
do
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
/*
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
* We need full exclusive lock on the relation in order to do
* truncation. If we can't get it, give up rather than waiting --- we
* don't want to block other backends, and we don't want to deadlock
* (which is quite possible considering we already hold a lower-grade
* lock).
Fix VACUUM so that it always updates pg_class.reltuples/relpages. When we added the ability for vacuum to skip heap pages by consulting the visibility map, we made it just not update the reltuples/relpages statistics if it skipped any pages. But this could leave us with extremely out-of-date stats for a table that contains any unchanging areas, especially for TOAST tables which never get processed by ANALYZE. In particular this could result in autovacuum making poor decisions about when to process the table, as in recent report from Florian Helmberger. And in general it's a bad idea to not update the stats at all. Instead, use the previous values of reltuples/relpages as an estimate of the tuple density in unvisited pages. This approach results in a "moving average" estimate of reltuples, which should converge to the correct value over multiple VACUUM and ANALYZE cycles even when individual measurements aren't very good. This new method for updating reltuples is used by both VACUUM and ANALYZE, with the result that we no longer need the grotty interconnections that caused ANALYZE to not update the stats depending on what had happened in the parent VACUUM command. Also, fix the logic for skipping all-visible pages during VACUUM so that it looks ahead rather than behind to decide what to do, as per a suggestion from Greg Stark. This eliminates useless scanning of all-visible pages at the start of the relation or just after a not-all-visible page. In particular, the first few pages of the relation will not be invariably included in the scanned pages, which seems to help in not overweighting them in the reltuples estimate. Back-patch to 8.4, where the visibility map was introduced.
2011-05-30 23:05:26 +02:00
*/
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
lock_waiter_detected = false;
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
lock_retry = 0;
while (true)
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (ConditionalLockRelation(vacrel->rel, AccessExclusiveLock))
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
break;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* Check for interrupts while trying to (re-)acquire the exclusive
* lock.
*/
CHECK_FOR_INTERRUPTS();
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
if (++lock_retry > (VACUUM_TRUNCATE_LOCK_TIMEOUT /
VACUUM_TRUNCATE_LOCK_WAIT_INTERVAL))
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
{
/*
* We failed to establish the lock in the specified number of
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
* retries. This means we give up truncating.
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
*/
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ereport(vacrel->verbose ? INFO : DEBUG2,
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
(errmsg("\"%s\": stopping truncate due to conflicting lock request",
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->relname)));
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
return;
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Use WaitLatch() instead of pg_usleep() at end-of-vacuum truncation This has the advantage to make a process more responsive when the postmaster dies, even if the wait time was rather limited as there was only a 50ms timeout here. Another advantage of this change is for monitoring, as we gain a new wait event for the end-of-vacuum truncation. Author: Bharath Rupireddy Reviewed-by: Aleksander Alekseev, Thomas Munro, Michael Paquier Discussion: https://postgr.es/m/CALj2ACU4AdPCq6NLfcA-ZGwX7pPCK5FgEj-CAU0xCKzkASSy_A@mail.gmail.com
2021-07-02 05:58:34 +02:00
(void) WaitLatch(MyLatch,
WL_LATCH_SET | WL_TIMEOUT | WL_EXIT_ON_PM_DEATH,
VACUUM_TRUNCATE_LOCK_WAIT_INTERVAL,
WAIT_EVENT_VACUUM_TRUNCATE);
ResetLatch(MyLatch);
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
}
Don't reset pg_class.reltuples and relpages in VACUUM, if any pages were skipped. We could update relpages anyway, but it seems better to only update it together with reltuples, because we use the reltuples/relpages ratio in the planner. Also don't update n_live_tuples in pgstat. ANALYZE in VACUUM ANALYZE now needs to update pg_class, if the VACUUM-phase didn't do so. Added some boolean-passing to let analyze_rel know if it should update pg_class or not. I also moved the relcache invalidation (to update rd_targblock) from vac_update_relstats to where RelationTruncate is called, because vac_update_relstats is not called for partial vacuums anymore. It's more obvious to send the invalidation close to the truncation that requires it. Per report by Ned T. Crigler.
2008-12-17 10:15:03 +01:00
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* Now that we have exclusive lock, look to see if the rel has grown
* whilst we were vacuuming with non-exclusive lock. If so, give up;
* the newly added pages presumably contain non-deletable tuples.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
new_rel_pages = RelationGetNumberOfBlocks(vacrel->rel);
Remove unneeded old_rel_pages VACUUM state field. The field hasn't been used since commit 3d351d91, which redefined pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Also rename a local variable of the same name ("old_rel_pages"). This is used by relation truncation to represent the original relation size at the start of the ongoing VACUUM operation. Rename it to orig_rel_pages, since that's a lot clearer. (This name matches similar nearby code.)
2021-08-31 23:59:52 +02:00
if (new_rel_pages != orig_rel_pages)
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
{
/*
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
* Note: we intentionally don't update vacrel->rel_pages with the
* new rel size here. If we did, it would amount to assuming that
* the new pages are empty, which is unlikely. Leaving the numbers
* alone amounts to assuming that the new pages have the same
* tuple density as existing ones, which is less unlikely.
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
UnlockRelation(vacrel->rel, AccessExclusiveLock);
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
return;
}
Release the exclusive lock on the table early after truncating it in lazy vacuum, instead of waiting till commit.
2007-09-10 23:40:03 +02:00
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* Scan backwards from the end to verify that the end pages actually
* contain no tuples. This is *necessary*, not optional, because
* other backends could have added tuples to these pages whilst we
* were vacuuming.
*/
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
new_rel_pages = count_nondeletable_pages(vacrel, &lock_waiter_detected);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->blkno = new_rel_pages;
Change planner to use the current true disk file size as its estimate of a relation's number of blocks, rather than the possibly-obsolete value in pg_class.relpages. Scale the value in pg_class.reltuples correspondingly to arrive at a hopefully more accurate number of rows. When pg_class contains 0/0, estimate a tuple width from the column datatypes and divide that into current file size to estimate number of rows. This improved methodology allows us to jettison the ancient hacks that put bogus default values into pg_class when a table is first created. Also, per a suggestion from Simon, make VACUUM (but not VACUUM FULL or ANALYZE) adjust the value it puts into pg_class.reltuples to try to represent the mean tuple density instead of the minimal density that actually prevails just after VACUUM. These changes alter the plans selected for certain regression tests, so update the expected files accordingly. (I removed join_1.out because it's not clear if it still applies; we can add back any variant versions as they are shown to be needed.)
2004-12-01 20:00:56 +01:00
Remove unneeded old_rel_pages VACUUM state field. The field hasn't been used since commit 3d351d91, which redefined pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Also rename a local variable of the same name ("old_rel_pages"). This is used by relation truncation to represent the original relation size at the start of the ongoing VACUUM operation. Rename it to orig_rel_pages, since that's a lot clearer. (This name matches similar nearby code.)
2021-08-31 23:59:52 +02:00
if (new_rel_pages >= orig_rel_pages)
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
{
/* can't do anything after all */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
UnlockRelation(vacrel->rel, AccessExclusiveLock);
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
return;
}
/*
* Okay to truncate.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
RelationTruncate(vacrel->rel, new_rel_pages);
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* We can release the exclusive lock as soon as we have truncated.
* Other backends can't safely access the relation until they have
* processed the smgr invalidation that smgrtruncate sent out ... but
* that should happen as part of standard invalidation processing once
* they acquire lock on the relation.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
UnlockRelation(vacrel->rel, AccessExclusiveLock);
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* Update statistics. Here, it *is* correct to adjust rel_pages
* without also touching reltuples, since the tuple count wasn't
* changed by the truncation.
*/
vacuumlazy.c: Rename state field for consistency. Rename pages_removed to removed_pages, for consistency with nearby vacrel fields.
2022-01-29 02:41:09 +01:00
vacrel->removed_pages += orig_rel_pages - new_rel_pages;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->rel_pages = new_rel_pages;
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ereport(vacrel->verbose ? INFO : DEBUG2,
VACUUM VERBOSE: Don't report "pages removed". It doesn't make any sense to report this information, since VACUUM VERBOSE reports on heap relation truncation directly. This was an oversight in commit 7ab96cf6, which made VACUUM VERBOSE output a little more consistent with nearby autovacuum-specific log output. Adjust comments that describe how this is supposed to work in passing. Also bring truncation-related VACUUM VERBOSE output in line with the convention established for VACUUM VERBOSE output by commit f4f4a649. Author: Peter Geoghegan <pg@bowt.ie> Backpatch: 14-, where VACUUM VERBOSE's output changed.
2021-09-01 05:37:18 +02:00
(errmsg("table \"%s\": truncated %u to %u pages",
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->relname,
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
orig_rel_pages, new_rel_pages)));
Remove unneeded old_rel_pages VACUUM state field. The field hasn't been used since commit 3d351d91, which redefined pg_class.reltuples to be -1 before the first VACUUM or ANALYZE. Also rename a local variable of the same name ("old_rel_pages"). This is used by relation truncation to represent the original relation size at the start of the ongoing VACUUM operation. Rename it to orig_rel_pages, since that's a lot clearer. (This name matches similar nearby code.)
2021-08-31 23:59:52 +02:00
orig_rel_pages = new_rel_pages;
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
} while (new_rel_pages > vacrel->nonempty_pages && lock_waiter_detected);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
/*
Fix aboriginal mistake in lazy VACUUM's code for truncating away no-longer-needed pages at the end of a table. We thought we could throw away pages containing HEAPTUPLE_DEAD tuples; but this is not so, because such tuples very likely have index entries pointing at them, and we wouldn't have removed the index entries. The problem only emerges in a somewhat unlikely race condition: the dead tuples have to have been inserted by a transaction that later aborted, and this has to have happened between VACUUM's initial scan of the page and then rechecking it for empty in count_nondeletable_pages. But that timespan will include an index-cleaning pass, so it's not all that hard to hit. This seems to explain a couple of previously unsolved bug reports.
2007-09-16 04:37:46 +02:00
* Rescan end pages to verify that they are (still) empty of tuples.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*
* Returns number of nondeletable pages (last nonempty page + 1).
*/
static BlockNumber
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
count_nondeletable_pages(LVRelState *vacrel, bool *lock_waiter_detected)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
BlockNumber blkno;
Prefetch blocks during lazy vacuum's truncation scan Vacuum truncation scan can be sped up on rotating media by prefetching blocks in forward direction. That makes the blocks already present in memory by the time they are needed, while also letting OS read-ahead kick in. The truncate scan has been measured to be five times faster than without this patch (that was on a slow disk, but it shouldn't hurt on fast disks.) Author: Álvaro Herrera, loosely based on a submission by Claudio Freire Discussion: https://postgr.es/m/CAGTBQpa6NFGO_6g_y_7zQx8L9GcHDSQKYdo1tGuh791z6PYgEg@mail.gmail.com
2017-01-23 16:55:18 +01:00
BlockNumber prefetchedUntil;
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
instr_time starttime;
/* Initialize the starttime if we check for conflicting lock requests */
INSTR_TIME_SET_CURRENT(starttime);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Prefetch blocks during lazy vacuum's truncation scan Vacuum truncation scan can be sped up on rotating media by prefetching blocks in forward direction. That makes the blocks already present in memory by the time they are needed, while also letting OS read-ahead kick in. The truncate scan has been measured to be five times faster than without this patch (that was on a slow disk, but it shouldn't hurt on fast disks.) Author: Álvaro Herrera, loosely based on a submission by Claudio Freire Discussion: https://postgr.es/m/CAGTBQpa6NFGO_6g_y_7zQx8L9GcHDSQKYdo1tGuh791z6PYgEg@mail.gmail.com
2017-01-23 16:55:18 +01:00
/*
* Start checking blocks at what we believe relation end to be and move
* backwards. (Strange coding of loop control is needed because blkno is
* unsigned.) To make the scan faster, we prefetch a few blocks at a time
* in forward direction, so that OS-level readahead can kick in.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
blkno = vacrel->rel_pages;
Prefetch blocks during lazy vacuum's truncation scan Vacuum truncation scan can be sped up on rotating media by prefetching blocks in forward direction. That makes the blocks already present in memory by the time they are needed, while also letting OS read-ahead kick in. The truncate scan has been measured to be five times faster than without this patch (that was on a slow disk, but it shouldn't hurt on fast disks.) Author: Álvaro Herrera, loosely based on a submission by Claudio Freire Discussion: https://postgr.es/m/CAGTBQpa6NFGO_6g_y_7zQx8L9GcHDSQKYdo1tGuh791z6PYgEg@mail.gmail.com
2017-01-23 16:55:18 +01:00
StaticAssertStmt((PREFETCH_SIZE & (PREFETCH_SIZE - 1)) == 0,
"prefetch size must be power of 2");
prefetchedUntil = InvalidBlockNumber;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
while (blkno > vacrel->nonempty_pages)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
Buffer buf;
Page page;
OffsetNumber offnum,
maxoff;
Fix aboriginal mistake in lazy VACUUM's code for truncating away no-longer-needed pages at the end of a table. We thought we could throw away pages containing HEAPTUPLE_DEAD tuples; but this is not so, because such tuples very likely have index entries pointing at them, and we wouldn't have removed the index entries. The problem only emerges in a somewhat unlikely race condition: the dead tuples have to have been inserted by a transaction that later aborted, and this has to have happened between VACUUM's initial scan of the page and then rechecking it for empty in count_nondeletable_pages. But that timespan will include an index-cleaning pass, so it's not all that hard to hit. This seems to explain a couple of previously unsolved bug reports.
2007-09-16 04:37:46 +02:00
bool hastup;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
/*
* Check if another process requests a lock on our relation. We are
* holding an AccessExclusiveLock here, so they will be waiting. We
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
* only do this once per VACUUM_TRUNCATE_LOCK_CHECK_INTERVAL, and we
* only check if that interval has elapsed once every 32 blocks to
* keep the number of system calls and actual shared lock table
* lookups to a minimum.
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
*/
if ((blkno % 32) == 0)
{
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
instr_time currenttime;
instr_time elapsed;
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
INSTR_TIME_SET_CURRENT(currenttime);
elapsed = currenttime;
INSTR_TIME_SUBTRACT(elapsed, starttime);
if ((INSTR_TIME_GET_MICROSEC(elapsed) / 1000)
Ensure ANALYZE phase is not skipped because of canceled truncate. Patch b19e4250b45e91c9cbdd18d35ea6391ab5961c8d attempted to preserve existing behavior regarding statistics generation in the case that a truncation attempt was canceled due to lock conflicts. It failed to do this accurately in two regards: (1) autovacuum had previously generated statistics if the truncate attempt failed to initially get the lock rather than having started the attempt, and (2) the VACUUM ANALYZE command had always generated statistics. Both of these changes were unintended, and are reverted by this patch. On review, there seems to be consensus that the previous failure to generate statistics when the truncate was terminated was more an unfortunate consequence of how that effort was previously terminated than a feature we want to keep; so this patch generates statistics even when an autovacuum truncation attempt terminates early. Another unintended change which is kept on the basis that it is an improvement is that when a VACUUM command is truncating, it will the new heuristic for avoiding blocking other processes, rather than keeping an AccessExclusiveLock on the table for however long the truncation takes. Per multiple reports, with some renaming per patch by Jeff Janes. Backpatch to 9.0, where problem was created.
2013-04-29 20:05:26 +02:00
>= VACUUM_TRUNCATE_LOCK_CHECK_INTERVAL)
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (LockHasWaitersRelation(vacrel->rel, AccessExclusiveLock))
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
{
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
ereport(vacrel->verbose ? INFO : DEBUG2,
VACUUM VERBOSE: Don't report "pages removed". It doesn't make any sense to report this information, since VACUUM VERBOSE reports on heap relation truncation directly. This was an oversight in commit 7ab96cf6, which made VACUUM VERBOSE output a little more consistent with nearby autovacuum-specific log output. Adjust comments that describe how this is supposed to work in passing. Also bring truncation-related VACUUM VERBOSE output in line with the convention established for VACUUM VERBOSE output by commit f4f4a649. Author: Peter Geoghegan <pg@bowt.ie> Backpatch: 14-, where VACUUM VERBOSE's output changed.
2021-09-01 05:37:18 +02:00
(errmsg("table \"%s\": suspending truncate due to conflicting lock request",
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->relname)));
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
Remove unneeded field from VACUUM state. Bugfix commit 5fc89376 effectively made the lock_waiter_detected field from vacuumlazy.c's global state struct into private state owned by lazy_truncate_heap(). Finish this off by replacing the struct field with a local variable.
2021-06-15 17:59:36 +02:00
*lock_waiter_detected = true;
Fix performance problems with autovacuum truncation in busy workloads. In situations where there are over 8MB of empty pages at the end of a table, the truncation work for trailing empty pages takes longer than deadlock_timeout, and there is frequent access to the table by processes other than autovacuum, there was a problem with the autovacuum worker process being canceled by the deadlock checking code. The truncation work done by autovacuum up that point was lost, and the attempt tried again by a later autovacuum worker. The attempts could continue indefinitely without making progress, consuming resources and blocking other processes for up to deadlock_timeout each time. This patch has the autovacuum worker checking whether it is blocking any other thread at 20ms intervals. If such a condition develops, the autovacuum worker will persist the work it has done so far, release its lock on the table, and sleep in 50ms intervals for up to 5 seconds, hoping to be able to re-acquire the lock and try again. If it is unable to get the lock in that time, it moves on and a worker will try to continue later from the point this one left off. While this patch doesn't change the rules about when and what to truncate, it does cause the truncation to occur sooner, with less blocking, and with the consumption of fewer resources when there is contention for the table's lock. The only user-visible change other than improved performance is that the table size during truncation may change incrementally instead of just once. This problem exists in all supported versions but is infrequently reported, although some reports of performance problems when autovacuum runs might be caused by this. Initial commit is just the master branch, but this should probably be backpatched once the build farm and general developer usage confirm that there are no surprising effects. Jan Wieck
2012-12-11 21:33:08 +01:00
return blkno;
}
starttime = currenttime;
}
}
Remove the vacuum_delay_point call in count_nondeletable_pages, because we hold an exclusive lock on the table at this point, which we want to release as soon as possible. This is called in the phase of lazy vacuum where we truncate the empty pages at the end of the table. An alternative solution would be to lower the vacuum delay settings before starting the truncating phase, but this doesn't work very well in autovacuum due to the autobalancing code (which can cause other processes to change our cost delay settings). This case could be considered in the balancing code, but it is simpler this way.
2007-09-10 19:58:45 +02:00
/*
* We don't insert a vacuum delay point here, because we have an
* exclusive lock on the table which we want to hold for as short a
Add a CHECK_FOR_INTERRUPTS call in the site where the vacuum delay point was removed.
2007-09-12 04:05:48 +02:00
* time as possible. We still need to check for interrupts however.
Remove the vacuum_delay_point call in count_nondeletable_pages, because we hold an exclusive lock on the table at this point, which we want to release as soon as possible. This is called in the phase of lazy vacuum where we truncate the empty pages at the end of the table. An alternative solution would be to lower the vacuum delay settings before starting the truncating phase, but this doesn't work very well in autovacuum due to the autobalancing code (which can cause other processes to change our cost delay settings). This case could be considered in the balancing code, but it is simpler this way.
2007-09-10 19:58:45 +02:00
*/
Add a CHECK_FOR_INTERRUPTS call in the site where the vacuum delay point was removed.
2007-09-12 04:05:48 +02:00
CHECK_FOR_INTERRUPTS();
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
blkno--;
Prefetch blocks during lazy vacuum's truncation scan Vacuum truncation scan can be sped up on rotating media by prefetching blocks in forward direction. That makes the blocks already present in memory by the time they are needed, while also letting OS read-ahead kick in. The truncate scan has been measured to be five times faster than without this patch (that was on a slow disk, but it shouldn't hurt on fast disks.) Author: Álvaro Herrera, loosely based on a submission by Claudio Freire Discussion: https://postgr.es/m/CAGTBQpa6NFGO_6g_y_7zQx8L9GcHDSQKYdo1tGuh791z6PYgEg@mail.gmail.com
2017-01-23 16:55:18 +01:00
/* If we haven't prefetched this lot yet, do so now. */
if (prefetchedUntil > blkno)
{
BlockNumber prefetchStart;
BlockNumber pblkno;
prefetchStart = blkno & ~(PREFETCH_SIZE - 1);
for (pblkno = prefetchStart; pblkno <= blkno; pblkno++)
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
PrefetchBuffer(vacrel->rel, MAIN_FORKNUM, pblkno);
Prefetch blocks during lazy vacuum's truncation scan Vacuum truncation scan can be sped up on rotating media by prefetching blocks in forward direction. That makes the blocks already present in memory by the time they are needed, while also letting OS read-ahead kick in. The truncate scan has been measured to be five times faster than without this patch (that was on a slow disk, but it shouldn't hurt on fast disks.) Author: Álvaro Herrera, loosely based on a submission by Claudio Freire Discussion: https://postgr.es/m/CAGTBQpa6NFGO_6g_y_7zQx8L9GcHDSQKYdo1tGuh791z6PYgEg@mail.gmail.com
2017-01-23 16:55:18 +01:00
CHECK_FOR_INTERRUPTS();
}
prefetchedUntil = prefetchStart;
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
buf = ReadBufferExtended(vacrel->rel, MAIN_FORKNUM, blkno, RBM_NORMAL,
vacrel->bstrategy);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/* In this phase we only need shared access to the buffer */
LockBuffer(buf, BUFFER_LOCK_SHARE);
Revert no-op changes to BufferGetPage() The reverted changes were intended to force a choice of whether any newly-added BufferGetPage() calls needed to be accompanied by a test of the snapshot age, to support the "snapshot too old" feature. Such an accompanying test is needed in about 7% of the cases, where the page is being used as part of a scan rather than positioning for other purposes (such as DML or vacuuming). The additional effort required for back-patching, and the doubt whether the intended benefit would really be there, have indicated it is best just to rely on developers to do the right thing based on comments and existing usage, as we do with many other conventions. This change should have little or no effect on generated executable code. Motivated by the back-patching pain of Tom Lane and Robert Haas
2016-04-20 15:31:19 +02:00
page = BufferGetPage(buf);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
if (PageIsNew(page) || PageIsEmpty(page))
{
Clean up WAL/buffer interactions as per my recent proposal. Get rid of the misleadingly-named WriteBuffer routine, and instead require routines that change buffer pages to call MarkBufferDirty (which does exactly what it says). We also require that they do so before calling XLogInsert; this takes care of the synchronization requirement documented in SyncOneBuffer. Note that because bufmgr takes the buffer content lock (in shared mode) while writing out any buffer, it doesn't matter whether MarkBufferDirty is executed before the buffer content change is complete, so long as the content change is completed before releasing exclusive lock on the buffer. So it's OK to set the dirtybit before we fill in the LSN. This eliminates the former kluge of needing to set the dirtybit in LockBuffer. Aside from making the code more transparent, we can also add some new debugging assertions, in particular that the caller of MarkBufferDirty must hold the buffer content lock, not merely a pin.
2006-04-01 01:32:07 +02:00
UnlockReleaseBuffer(buf);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
continue;
}
hastup = false;
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber;
offnum <= maxoff;
offnum = OffsetNumberNext(offnum))
{
ItemId itemid;
itemid = PageGetItemId(page, offnum);
Fix aboriginal mistake in lazy VACUUM's code for truncating away no-longer-needed pages at the end of a table. We thought we could throw away pages containing HEAPTUPLE_DEAD tuples; but this is not so, because such tuples very likely have index entries pointing at them, and we wouldn't have removed the index entries. The problem only emerges in a somewhat unlikely race condition: the dead tuples have to have been inserted by a transaction that later aborted, and this has to have happened between VACUUM's initial scan of the page and then rechecking it for empty in count_nondeletable_pages. But that timespan will include an index-cleaning pass, so it's not all that hard to hit. This seems to explain a couple of previously unsolved bug reports.
2007-09-16 04:37:46 +02:00
/*
* Note: any non-unused item should be taken as a reason to keep
Support disabling index bypassing by VACUUM. Generalize the INDEX_CLEANUP VACUUM parameter (and the corresponding reloption): make it into a ternary style boolean parameter. It now exposes a third option, "auto". The "auto" option (which is now the default) enables the "bypass index vacuuming" optimization added by commit 1e55e7d1. "VACUUM (INDEX_CLEANUP TRUE)" is redefined to once again make VACUUM simply do any required index vacuuming, regardless of how few dead tuples are encountered during the first scan of the target heap relation (unless there are exactly zero). This gives users a way of opting out of the "bypass index vacuuming" optimization, if for whatever reason that proves necessary. It is also expected to be used by PostgreSQL developers as a testing option from time to time. "VACUUM (INDEX_CLEANUP FALSE)" does the same thing as it always has: it forcibly disables both index vacuuming and index cleanup. It's not expected to be used much in PostgreSQL 14. The failsafe mechanism added by commit 1e55e7d1 addresses the same problem in a simpler way. INDEX_CLEANUP can now be thought of as a testing and compatibility option. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Justin Pryzby <pryzby@telsasoft.com> Discussion: https://postgr.es/m/CAH2-WznrBoCST4_Gxh_G9hA8NzGUbeBGnOUC8FcXcrhqsv6OHQ@mail.gmail.com
2021-06-19 05:04:07 +02:00
* this page. Even an LP_DEAD item makes truncation unsafe, since
* we must not have cleaned out its index entries.
Fix aboriginal mistake in lazy VACUUM's code for truncating away no-longer-needed pages at the end of a table. We thought we could throw away pages containing HEAPTUPLE_DEAD tuples; but this is not so, because such tuples very likely have index entries pointing at them, and we wouldn't have removed the index entries. The problem only emerges in a somewhat unlikely race condition: the dead tuples have to have been inserted by a transaction that later aborted, and this has to have happened between VACUUM's initial scan of the page and then rechecking it for empty in count_nondeletable_pages. But that timespan will include an index-cleaning pass, so it's not all that hard to hit. This seems to explain a couple of previously unsolved bug reports.
2007-09-16 04:37:46 +02:00
*/
if (ItemIdIsUsed(itemid))
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
hastup = true;
break; /* can stop scanning */
}
} /* scan along page */
Clean up WAL/buffer interactions as per my recent proposal. Get rid of the misleadingly-named WriteBuffer routine, and instead require routines that change buffer pages to call MarkBufferDirty (which does exactly what it says). We also require that they do so before calling XLogInsert; this takes care of the synchronization requirement documented in SyncOneBuffer. Note that because bufmgr takes the buffer content lock (in shared mode) while writing out any buffer, it doesn't matter whether MarkBufferDirty is executed before the buffer content change is complete, so long as the content change is completed before releasing exclusive lock on the buffer. So it's OK to set the dirtybit before we fill in the LSN. This eliminates the former kluge of needing to set the dirtybit in LockBuffer. Aside from making the code more transparent, we can also add some new debugging assertions, in particular that the caller of MarkBufferDirty must hold the buffer content lock, not merely a pin.
2006-04-01 01:32:07 +02:00
UnlockReleaseBuffer(buf);
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
/* Done scanning if we found a tuple here */
if (hastup)
return blkno + 1;
}
/*
* If we fall out of the loop, all the previously-thought-to-be-empty
Fix aboriginal mistake in lazy VACUUM's code for truncating away no-longer-needed pages at the end of a table. We thought we could throw away pages containing HEAPTUPLE_DEAD tuples; but this is not so, because such tuples very likely have index entries pointing at them, and we wouldn't have removed the index entries. The problem only emerges in a somewhat unlikely race condition: the dead tuples have to have been inserted by a transaction that later aborted, and this has to have happened between VACUUM's initial scan of the page and then rechecking it for empty in count_nondeletable_pages. But that timespan will include an index-cleaning pass, so it's not all that hard to hit. This seems to explain a couple of previously unsolved bug reports.
2007-09-16 04:37:46 +02:00
* pages still are; we need not bother to look at the last known-nonempty
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
* page.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
return vacrel->nonempty_pages;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
/*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* Returns the number of dead TIDs that VACUUM should allocate space to
* store, given a heap rel of size vacrel->rel_pages, and given current
* maintenance_work_mem setting (or current autovacuum_work_mem setting,
* when applicable).
*
* See the comments at the head of this file for rationale.
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
*/
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
static int
dead_items_max_items(LVRelState *vacrel)
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
{
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
int64 max_items;
New autovacuum_work_mem parameter If autovacuum_work_mem is set, autovacuum workers now use this parameter in preference to maintenance_work_mem. Peter Geoghegan
2013-12-12 12:42:39 +01:00
int vac_work_mem = IsAutoVacuumWorkerProcess() &&
Pre-beta mechanical code beautification. Run pgindent, pgperltidy, and reformat-dat-files. This set of diffs is a bit larger than typical. We've updated to pg_bsd_indent 2.1.2, which properly indents variable declarations that have multi-line initialization expressions (the continuation lines are now indented one tab stop). We've also updated to perltidy version 20230309 and changed some of its settings, which reduces its desire to add whitespace to lines to make assignments etc. line up. Going forward, that should make for fewer random-seeming changes to existing code. Discussion: https://postgr.es/m/20230428092545.qfb3y5wcu4cm75ur@alvherre.pgsql
2023-05-19 23:24:48 +02:00
autovacuum_work_mem != -1 ?
autovacuum_work_mem : maintenance_work_mem;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
if (vacrel->nindexes > 0)
Code review for patch to avoid second scan when vacuuming index-less table: avoid invoking LockBufferForCleanup without need, put out the same log message we would have before, minor code beautification.
2006-09-13 19:47:08 +02:00
{
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
BlockNumber rel_pages = vacrel->rel_pages;
max_items = MAXDEADITEMS(vac_work_mem * 1024L);
max_items = Min(max_items, INT_MAX);
max_items = Min(max_items, MAXDEADITEMS(MaxAllocSize));
Reduce the size of memory allocations by lazy vacuum when processing a small table, by allocating just enough for a hardcoded number of dead tuples per page. The current estimate is 200 dead tuples per page. Per reports from Jeff Amiel, Erik Jones and Marko Kreen, and subsequent discussion. CVS: ---------------------------------------------------------------------- CVS: Enter Log. Lines beginning with `CVS:' are removed automatically CVS: CVS: Committing in . CVS: CVS: Modified Files: CVS: commands/vacuumlazy.c CVS: ----------------------------------------------------------------------
2007-09-24 05:52:55 +02:00
/* curious coding here to ensure the multiplication can't overflow */
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
if ((BlockNumber) (max_items / MaxHeapTuplesPerPage) > rel_pages)
max_items = rel_pages * MaxHeapTuplesPerPage;
Reduce the size of memory allocations by lazy vacuum when processing a small table, by allocating just enough for a hardcoded number of dead tuples per page. The current estimate is 200 dead tuples per page. Per reports from Jeff Amiel, Erik Jones and Marko Kreen, and subsequent discussion. CVS: ---------------------------------------------------------------------- CVS: Enter Log. Lines beginning with `CVS:' are removed automatically CVS: CVS: Committing in . CVS: CVS: Modified Files: CVS: commands/vacuumlazy.c CVS: ----------------------------------------------------------------------
2007-09-24 05:52:55 +02:00
Trivial patch to double vacuum speed on tables with no indexes (prevent second scan of table). Gregory Stark
2006-09-04 23:40:23 +02:00
/* stay sane if small maintenance_work_mem */
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
max_items = Max(max_items, MaxHeapTuplesPerPage);
Code review for patch to avoid second scan when vacuuming index-less table: avoid invoking LockBufferForCleanup without need, put out the same log message we would have before, minor code beautification.
2006-09-13 19:47:08 +02:00
}
else
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
{
/* One-pass case only stores a single heap page's TIDs at a time */
max_items = MaxHeapTuplesPerPage;
}
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
return (int) max_items;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
}
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
/*
* Allocate dead_items (either using palloc, or in dynamic shared memory).
* Sets dead_items in vacrel for caller.
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* Also handles parallel initialization as part of allocating dead_items in
* DSM when required.
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
*/
static void
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
dead_items_alloc(LVRelState *vacrel, int nworkers)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
{
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
VacDeadItems *dead_items;
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
int max_items;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
max_items = dead_items_max_items(vacrel);
Assert(max_items >= MaxHeapTuplesPerPage);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
/*
* Initialize state for a parallel vacuum. As of now, only one worker can
* be used for an index, so we invoke parallelism only if there are at
* least two indexes on a table.
*/
Add wraparound failsafe to VACUUM. Add a failsafe mechanism that is triggered by VACUUM when it notices that the table's relfrozenxid and/or relminmxid are dangerously far in the past. VACUUM checks the age of the table dynamically, at regular intervals. When the failsafe triggers, VACUUM takes extraordinary measures to finish as quickly as possible so that relfrozenxid and/or relminmxid can be advanced. VACUUM will stop applying any cost-based delay that may be in effect. VACUUM will also bypass any further index vacuuming and heap vacuuming -- it only completes whatever remaining pruning and freezing is required. Bypassing index/heap vacuuming is enabled by commit 8523492d, which made it possible to dynamically trigger the mechanism already used within VACUUM when it is run with INDEX_CLEANUP off. It is expected that the failsafe will almost always trigger within an autovacuum to prevent wraparound, long after the autovacuum began. However, the failsafe mechanism can trigger in any VACUUM operation. Even in a non-aggressive VACUUM, where we're likely to not advance relfrozenxid, it still seems like a good idea to finish off remaining pruning and freezing. An aggressive/anti-wraparound VACUUM will be launched immediately afterwards. Note that the anti-wraparound VACUUM that follows will itself trigger the failsafe, usually before it even begins its first (and only) pass over the heap. The failsafe is controlled by two new GUCs: vacuum_failsafe_age, and vacuum_multixact_failsafe_age. There are no equivalent reloptions, since that isn't expected to be useful. The GUCs have rather high defaults (both default to 1.6 billion), and are expected to generally only be used to make the failsafe trigger sooner/more frequently. Author: Masahiko Sawada <sawada.mshk@gmail.com> Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAD21AoD0SkE11fMw4jD4RENAwBMcw1wasVnwpJVw3tVqPOQgAw@mail.gmail.com Discussion: https://postgr.es/m/CAH2-WzmgH3ySGYeC-m-eOBsa2=sDwa292-CFghV4rESYo39FsQ@mail.gmail.com
2021-04-07 21:37:45 +02:00
if (nworkers >= 0 && vacrel->nindexes > 1 && vacrel->do_index_vacuuming)
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
{
/*
* Since parallel workers cannot access data in temporary tables, we
* can't perform parallel vacuum on them.
*/
if (RelationUsesLocalBuffers(vacrel->rel))
{
/*
* Give warning only if the user explicitly tries to perform a
* parallel vacuum on the temporary table.
*/
if (nworkers > 0)
ereport(WARNING,
(errmsg("disabling parallel option of vacuum on \"%s\" --- cannot vacuum temporary tables in parallel",
vacrel->relname)));
}
else
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
vacrel->pvs = parallel_vacuum_init(vacrel->rel, vacrel->indrels,
vacrel->nindexes, nworkers,
Unify VACUUM VERBOSE and autovacuum logging. The log_autovacuum_min_duration instrumentation used its own dedicated code for logging, which was not reused by VACUUM VERBOSE. This was highly duplicative, and sometimes led to each code path using slightly different accounting for essentially the same information. Clean things up by making VACUUM VERBOSE reuse the same instrumentation code. This code restructuring changes the structure of the VACUUM VERBOSE output itself, but that seems like an overall improvement. The most noticeable change in VACUUM VERBOSE output is that it no longer outputs a distinct message per index per round of index vacuuming. Most of the same information (about each index) is now shown in its new per-operation summary message. This is far more legible. A few details are no longer displayed by VACUUM VERBOSE, but that's no real loss in practice, especially in the common case where we don't need multiple index scans/rounds of vacuuming. This super fine-grained information is still available via DEBUG2 messages, which might still be useful in debugging scenarios. VACUUM VERBOSE now shows new instrumentation, which is typically very useful: all of the log_autovacuum_min_duration instrumentation that it missed out on before now. This includes information about WAL overhead, buffers hit/missed/dirtied information, and I/O timing information. VACUUM VERBOSE still retains a few INFO messages of its own. This is limited to output concerning the progress of heap rel truncation, as well as some basic information about parallel workers. These details are still potentially quite useful. They aren't a good fit for the log output, which must summarize the whole operation. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-WzmW4Me7_qR4X4ka7pxP-jGmn7=Npma_-Z-9Y1eD0MQRLw@mail.gmail.com
2022-01-15 01:50:34 +01:00
max_items,
vacrel->verbose ? INFO : DEBUG2,
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
vacrel->bstrategy);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
/* If parallel mode started, dead_items space is allocated in DSM */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (ParallelVacuumIsActive(vacrel))
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
{
vacrel->dead_items = parallel_vacuum_get_dead_items(vacrel->pvs);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
return;
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
}
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
/* Serial VACUUM case */
Move index vacuum routines to vacuum.c. An upcoming patch moves parallel vacuum code out of vacuumlazy.c. This code restructuring will allow both lazy vacuum and parallel vacuum to use index vacuum functions. Author: Masahiko Sawada Reviewed-by: Hou Zhijie, Amit Kapila Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-22 03:25:14 +01:00
dead_items = (VacDeadItems *) palloc(vac_max_items_to_alloc_size(max_items));
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
dead_items->max_items = max_items;
dead_items->num_items = 0;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
vacrel->dead_items = dead_items;
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
}
/*
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
* Perform cleanup for resources allocated in dead_items_alloc
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
*/
static void
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
dead_items_cleanup(LVRelState *vacrel)
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
{
if (!ParallelVacuumIsActive(vacrel))
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
{
/* Don't bother with pfree here */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
return;
vacuumlazy.c: Rename dead_tuples to dead_items. Commit 8523492d simplified what it meant for an item to be considered "dead" to VACUUM: TIDs collected in memory (in preparation for index vacuuming) must always come from LP_DEAD stub line pointers in heap pages, found following pruning. This formalized the idea that index vacuuming (and heap vacuuming) are optional processes. Unlike pruning, they can be delayed indefinitely, without any risk of that violating fundamental invariants. For example, leaving LP_DEAD items behind clearly won't add to the risk of transaction ID wraparound. You can't have transaction ID wraparound without transaction IDs. Renaming anything that references DEAD tuples (tuples with storage) reinforces all this. Code outside vacuumlazy.c continues to fudge the distinction between dead/deleted tuples, and LP_DEAD items. This is necessary because autovacuum scheduling is still mostly driven by "dead items/tuples" statistics. In the future we may find it useful to replace this model with something more sophisticated, as a step towards teaching autovacuum to perform more frequent vacuuming that targeting individual indexes that happen to be more prone to becoming bloated through version churn. In passing, simplify some function signatures that deal with VACUUM's dead_items array. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Discussion: https://postgr.es/m/CAH2-WzktGBg4si6DEdmq3q6SoXSDqNi6MtmB8CmmTmvhsxDTLA@mail.gmail.com
2021-11-29 18:58:01 +01:00
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
/* End parallel mode */
parallel_vacuum_end(vacrel->pvs, vacrel->indstats);
vacrel->pvs = NULL;
Initial implementation of concurrent VACUUM. Ifdef'd out for the moment, because index locking issues are not handled correctly yet. Need to go work on the index AMs next.
2001-07-14 00:55:59 +02:00
}
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
/*
* Check if every tuple in the given page is visible to all current and future
* transactions. Also return the visibility_cutoff_xid which is the highest
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
* xmin amongst the visible tuples. Set *all_frozen to true if every tuple
* on this page is frozen.
Update another obsolete reference in vacuumlazy.c. Addresses an oversight in commit 7ab96cf6.
2021-11-11 22:13:08 +01:00
*
* This is a stripped down version of lazy_scan_prune(). If you change
* anything here, make sure that everything stays in sync. Note that an
* assertion calls us to verify that everybody still agrees. Be sure to avoid
* introducing new side-effects here.
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
*/
static bool
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
heap_page_is_all_visible(LVRelState *vacrel, Buffer buf,
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
TransactionId *visibility_cutoff_xid,
bool *all_frozen)
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
{
Revert no-op changes to BufferGetPage() The reverted changes were intended to force a choice of whether any newly-added BufferGetPage() calls needed to be accompanied by a test of the snapshot age, to support the "snapshot too old" feature. Such an accompanying test is needed in about 7% of the cases, where the page is being used as part of a scan rather than positioning for other purposes (such as DML or vacuuming). The additional effort required for back-patching, and the doubt whether the intended benefit would really be there, have indicated it is best just to rely on developers to do the right thing based on comments and existing usage, as we do with many other conventions. This change should have little or no effect on generated executable code. Motivated by the back-patching pain of Tom Lane and Robert Haas
2016-04-20 15:31:19 +02:00
Page page = BufferGetPage(buf);
Move BufferGetBlockNumber() out of heap_page_is_all_visible()'s inner loop. In some workloads BufferGetBlockNumber() shows up in profiles due to the sheer number of calls to it (and because it causes cache misses). The compiler can't move it out of the loop because it's a full extern function call...
2014-11-14 17:04:44 +01:00
BlockNumber blockno = BufferGetBlockNumber(buf);
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
OffsetNumber offnum,
maxoff;
bool all_visible = true;
*visibility_cutoff_xid = InvalidTransactionId;
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
*all_frozen = true;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
maxoff = PageGetMaxOffsetNumber(page);
for (offnum = FirstOffsetNumber;
offnum <= maxoff && all_visible;
offnum = OffsetNumberNext(offnum))
{
ItemId itemid;
HeapTupleData tuple;
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
/*
* Set the offset number so that we can display it along with any
* error that occurred while processing this tuple.
*/
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->offnum = offnum;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
itemid = PageGetItemId(page, offnum);
/* Unused or redirect line pointers are of no interest */
if (!ItemIdIsUsed(itemid) || ItemIdIsRedirected(itemid))
continue;
Move BufferGetBlockNumber() out of heap_page_is_all_visible()'s inner loop. In some workloads BufferGetBlockNumber() shows up in profiles due to the sheer number of calls to it (and because it causes cache misses). The compiler can't move it out of the loop because it's a full extern function call...
2014-11-14 17:04:44 +01:00
ItemPointerSet(&(tuple.t_self), blockno, offnum);
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
/*
* Dead line pointers can have index pointers pointing to them. So
* they can't be treated as visible
*/
if (ItemIdIsDead(itemid))
{
all_visible = false;
Cosmetic improvements to freeze map code. Per post-commit review comments from Andres Freund, improve variable names, comments, and in one place, slightly improve the code structure. Masahiko Sawada
2016-06-03 14:43:41 +02:00
*all_frozen = false;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
break;
}
Assert(ItemIdIsNormal(itemid));
tuple.t_data = (HeapTupleHeader) PageGetItem(page, itemid);
Adjust HeapTupleSatisfies* routines to take a HeapTuple. Previously, these functions took a HeapTupleHeader, but upcoming patches for logical replication will introduce new a new snapshot type under which the tuple's TID will be used to lookup (CMIN, CMAX) for visibility determination purposes. This makes that information available. Code churn is minimal since HeapTupleSatisfiesVisibility took the HeapTuple anyway, and deferenced it before calling the satisfies function. Independently of logical replication, this allows t_tableOid and t_self to be cross-checked via assertions in tqual.c. This seems like a useful way to make sure that all callers are setting these values properly, which has been previously put forward as desirable. Andres Freund, reviewed by Álvaro Herrera
2013-07-22 19:26:33 +02:00
tuple.t_len = ItemIdGetLength(itemid);
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
tuple.t_tableOid = RelationGetRelid(vacrel->rel);
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
switch (HeapTupleSatisfiesVacuum(&tuple, vacrel->cutoffs.OldestXmin,
buf))
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
{
case HEAPTUPLE_LIVE:
{
TransactionId xmin;
Update obsolete reference in vacuumlazy.c. Oversight in commit 7ab96cf6.
2021-11-06 07:38:07 +01:00
/* Check comments in lazy_scan_prune. */
Change the way we mark tuples as frozen. Instead of changing the tuple xmin to FrozenTransactionId, the combination of HEAP_XMIN_COMMITTED and HEAP_XMIN_INVALID, which were previously never set together, is now defined as HEAP_XMIN_FROZEN. A variety of previous proposals to freeze tuples opportunistically before vacuum_freeze_min_age is reached have foundered on the objection that replacing xmin by FrozenTransactionId might hinder debugging efforts when things in this area go awry; this patch is intended to solve that problem by keeping the XID around (but largely ignoring the value to which it is set). Third-party code that checks for HEAP_XMIN_INVALID on tuples where HEAP_XMIN_COMMITTED might be set will be broken by this change. To fix, use the new accessor macros in htup_details.h rather than consulting the bits directly. HeapTupleHeaderGetXmin has been modified to return FrozenTransactionId when the infomask bits indicate that the tuple is frozen; use HeapTupleHeaderGetRawXmin when you already know that the tuple isn't marked commited or frozen, or want the raw value anyway. We currently do this in routines that display the xmin for user consumption, in tqual.c where it's known to be safe and important for the avoidance of extra cycles, and in the function-caching code for various procedural languages, which shouldn't invalidate the cache just because the tuple gets frozen. Robert Haas and Andres Freund
2013-12-22 21:49:09 +01:00
if (!HeapTupleHeaderXminCommitted(tuple.t_data))
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
{
all_visible = false;
Cosmetic improvements to freeze map code. Per post-commit review comments from Andres Freund, improve variable names, comments, and in one place, slightly improve the code structure. Masahiko Sawada
2016-06-03 14:43:41 +02:00
*all_frozen = false;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
break;
}
/*
* The inserter definitely committed. But is it old enough
* that everyone sees it as committed?
*/
xmin = HeapTupleHeaderGetXmin(tuple.t_data);
Refactor how VACUUM passes around its XID cutoffs. Use a dedicated struct for the XID/MXID cutoffs used by VACUUM, such as FreezeLimit and OldestXmin. This state is initialized in vacuum.c, and then passed around by code from vacuumlazy.c to heapam.c freezing related routines. The new convention is that everybody works off of the same cutoff state, which is passed around via pointers to const. Also simplify some of the logic for dealing with frozen xmin in heap_prepare_freeze_tuple: add dedicated "xmin_already_frozen" state to clearly distinguish xmin XIDs that we're going to freeze from those that were already frozen from before. That way the routine's xmin handling code is symmetrical with the existing xmax handling code. This is preparation for an upcoming commit that will add page level freezing. Also refactor the control flow within FreezeMultiXactId(), while adding stricter sanity checks. We now test OldestXmin directly, instead of using FreezeLimit as an inexact proxy for OldestXmin. This is further preparation for the page level freezing work, which will make the function's caller cede control of page level freezing to the function where appropriate (where heap_prepare_freeze_tuple sees a tuple that happens to contain a MultiXactId in its xmax). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Jeff Davis <pgsql@j-davis.com> Discussion: https://postgr.es/m/CAH2-WznS9TxXmz2_=SY+SyJyDFbiOftKofM9=aDo68BbXNBUMA@mail.gmail.com
2022-12-22 18:37:59 +01:00
if (!TransactionIdPrecedes(xmin,
vacrel->cutoffs.OldestXmin))
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
{
all_visible = false;
Cosmetic improvements to freeze map code. Per post-commit review comments from Andres Freund, improve variable names, comments, and in one place, slightly improve the code structure. Masahiko Sawada
2016-06-03 14:43:41 +02:00
*all_frozen = false;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
break;
}
/* Track newest xmin on page. */
Avoid special XID snapshotConflictHorizon values. Don't allow VACUUM to WAL-log the value FrozenTransactionId as the snapshotConflictHorizon of freezing or visibility map related WAL records. The only special XID value that's an allowable snapshotConflictHorizon is InvalidTransactionId, which is interpreted as "record definitely doesn't require a recovery conflict". Author: Peter Geoghegan <pg@bowt.ie> Discussion: https://postgr.es/m/CAH2-WznuNGSzF8v6OsgjaC5aYsb3cZ6HW6MLm30X0d65cmSH6A@mail.gmail.com
2023-01-02 19:16:51 +01:00
if (TransactionIdFollows(xmin, *visibility_cutoff_xid) &&
TransactionIdIsNormal(xmin))
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
*visibility_cutoff_xid = xmin;
Change the format of the VM fork to add a second bit per page. The new bit indicates whether every tuple on the page is already frozen. It is cleared only when the all-visible bit is cleared, and it can be set only when we vacuum a page and find that every tuple on that page is both visible to every transaction and in no need of any future vacuuming. A future commit will use this new bit to optimize away full-table scans that would otherwise be triggered by XID wraparound considerations. A page which is merely all-visible must still be scanned in that case, but a page which is all-frozen need not be. This commit does not attempt that optimization, although that optimization is the goal here. It seems better to get the basic infrastructure in place first. Per discussion, it's very desirable for pg_upgrade to automatically migrate existing VM forks from the old format to the new format. That, too, will be handled in a follow-on patch. Masahiko Sawada, reviewed by Kyotaro Horiguchi, Fujii Masao, Amit Kapila, Simon Riggs, Andres Freund, and others, and substantially revised by me.
2016-03-02 03:49:41 +01:00
/* Check whether this tuple is already frozen or not */
if (all_visible && *all_frozen &&
heap_tuple_needs_eventual_freeze(tuple.t_data))
*all_frozen = false;
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
}
break;
case HEAPTUPLE_DEAD:
case HEAPTUPLE_RECENTLY_DEAD:
case HEAPTUPLE_INSERT_IN_PROGRESS:
case HEAPTUPLE_DELETE_IN_PROGRESS:
Cosmetic improvements to freeze map code. Per post-commit review comments from Andres Freund, improve variable names, comments, and in one place, slightly improve the code structure. Masahiko Sawada
2016-06-03 14:43:41 +02:00
{
all_visible = false;
*all_frozen = false;
break;
}
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
default:
elog(ERROR, "unexpected HeapTupleSatisfiesVacuum result");
break;
}
} /* scan along page */
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
/* Clear the offset information once we have processed the given page. */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->offnum = InvalidOffsetNumber;
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
Update visibility map in the second phase of vacuum. There's a high chance that a page becomes all-visible when the second phase of vacuum removes all the dead tuples on it, so it makes sense to check for that. Otherwise the visibility map won't get updated until the next vacuum. Pavan Deolasee, reviewed by Jeff Janes.
2013-02-13 16:46:23 +01:00
return all_visible;
}
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
/*
Comments and doc fixes for commit 40d964ec99. Reported-by: Justin Pryzby Author: Justin Pryzby, with few changes by me Reviewed-by: Amit Kapila and Sawada Masahiko Discussion: https://postgr.es/m/20200322021801.GB2563@telsasoft.com
2020-04-14 04:40:27 +02:00
* Update index statistics in pg_class if the statistics are accurate.
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
*/
static void
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
update_relstats_all_indexes(LVRelState *vacrel)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Relation *indrels = vacrel->indrels;
int nindexes = vacrel->nindexes;
IndexBulkDeleteResult **indstats = vacrel->indstats;
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
vacuumlazy.c: document vistest and OldestXmin. Explain the relationship between vacuumlazy.c's vistest and OldestXmin cutoffs. These closely related cutoffs are different in subtle but important ways. Also document a closely related rule: we must establish rel_pages _after_ OldestXmin to ensure that no XID < OldestXmin can be missed by lazy_scan_heap(). It's easier to explain these issues by initializing everything together, so consolidate initialization of vacrel state. Now almost every vacrel field is initialized by heap_vacuum_rel(). The only remaining exception is the dead_items array, which is still managed by lazy_scan_heap() due to interactions with how we initialize parallel VACUUM. Also move the process that updates pg_class entries for each index into heap_vacuum_rel(), and adjust related assertions. All pg_class updates now take place after lazy_scan_heap() returns, which seems clearer. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/20211211045710.ljtuu4gfloh754rs@alap3.anarazel.de Discussion: https://postgr.es/m/CAH2-WznYsUxVT156rCQ+q=YD4S4=1M37hWvvHLz-H1pwSM8-Ew@mail.gmail.com
2022-03-12 21:52:38 +01:00
Assert(vacrel->do_index_cleanup);
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
for (int idx = 0; idx < nindexes; idx++)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
Relation indrel = indrels[idx];
IndexBulkDeleteResult *istat = indstats[idx];
if (istat == NULL || istat->estimated_count)
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
continue;
/* Update index statistics */
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vac_update_relstats(indrel,
istat->num_pages,
istat->num_index_tuples,
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
0,
false,
InvalidTransactionId,
InvalidMultiXactId,
Add VACUUM instrumentation for scanned pages, relfrozenxid. Report on scanned pages within VACUUM VERBOSE and autovacuum logging. These are pages that were physically examined during the VACUUM operation. Note that this can include a small number of pages that were marked all-visible in the visibility map by some earlier VACUUM operation. VACUUM won't skip all-visible pages that aren't part of a range of all-visible pages that's at least 32 blocks in length (partly to avoid missing out on opportunities to advance relfrozenxid during non-aggressive VACUUMs). Commit 44fa8488 simplified the definition of scanned pages. It became the complement of the pages (of those pages from rel_pages) that were skipped using the visibility map. And so scanned pages precisely indicates how effective the visibility map was at saving work. (Before now we displayed the number of pages skipped via the visibility map when happened to be frozen pages, but not when they were merely all-visible, which was less useful to users.) Rename the user-visible OldestXmin output field to "removal cutoff", and show some supplementary information: how far behind the cutoff is (number of XIDs behind) by the time the VACUUM operation finished. This will help users to figure out what's _not_ working in extreme cases where VACUUM is fundamentally unable to remove dead tuples or freeze older tuples (e.g., due to a leaked replication slot). Also report when relfrozenxid is advanced by VACUUM in output that immediately follows "removal cutoff". This structure is intended to highlight the relationship between the new relfrozenxid value for the table, and the VACUUM operation's removal cutoff. Finally, add instrumentation of "missed dead tuples", and the number of pages that had at least one such tuple. These are fully DEAD (not just RECENTLY_DEAD) tuples with storage that could not be pruned due to failure to acquire a cleanup lock on a heap page. This is a replacement for the "skipped due to pin" instrumentation removed by commit 44fa8488. It shows more details than before for pages where failing to get a cleanup lock actually resulted in VACUUM missing out on useful work, but usually shows nothing at all instead (the mere fact that we couldn't get a cleanup lock is usually of no consequence whatsoever now). Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Andres Freund <andres@anarazel.de> Discussion: https://postgr.es/m/CAH2-Wznp=c=Opj8Z7RMR3G=ec3_JfGYMN_YvmCEjoPCHzWbx0g@mail.gmail.com
2022-02-12 01:48:40 +01:00
NULL, NULL, false);
Allow vacuum command to process indexes in parallel. This feature allows the vacuum to leverage multiple CPUs in order to process indexes. This enables us to perform index vacuuming and index cleanup with background workers. This adds a PARALLEL option to VACUUM command where the user can specify the number of workers that can be used to perform the command which is limited by the number of indexes on a table. Specifying zero as a number of workers will disable parallelism. This option can't be used with the FULL option. Each index is processed by at most one vacuum process. Therefore parallel vacuum can be used when the table has at least two indexes. The parallel degree is either specified by the user or determined based on the number of indexes that the table has, and further limited by max_parallel_maintenance_workers. The index can participate in parallel vacuum iff it's size is greater than min_parallel_index_scan_size. Author: Masahiko Sawada and Amit Kapila Reviewed-by: Dilip Kumar, Amit Kapila, Robert Haas, Tomas Vondra, Mahendra Singh and Sergei Kornilov Tested-by: Mahendra Singh and Prabhat Sahu Discussion: https://postgr.es/m/CAD21AoDTPMgzSkV4E3SFo1CH_x50bf5PqZFQf4jmqjk-C03BWg@mail.gmail.com https://postgr.es/m/CAA4eK1J-VoR9gzS5E75pcD-OH0mEyCdp8RihcwKrcuw7J-Q0+w@mail.gmail.com
2020-01-20 03:27:49 +01:00
}
}
/*
Move parallel vacuum code to vacuumparallel.c. This commit moves parallel vacuum related code to a new file commands/vacuumparallel.c so that any table AM supporting indexes can utilize parallel vacuum in order to call index AM callbacks (ambulkdelete and amvacuumcleanup) with parallel workers. Another reason for this refactoring is that the parallel vacuum isn't specific to heap so it doesn't make sense to keep this code in heap/vacuumlazy.c. Author: Masahiko Sawada, based on suggestion from Andres Freund Reviewed-by: Hou Zhijie, Amit Kapila, Haiying Tang Discussion: https://www.postgresql.org/message-id/20211030212101.ae3qcouatwmy7tbr%40alap3.anarazel.de
2021-12-23 07:12:52 +01:00
* Error context callback for errors occurring during vacuum. The error
* context messages for index phases should match the messages set in parallel
* vacuum. If you change this function for those phases, change
* parallel_vacuum_error_callback() as well.
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
*/
static void
vacuum_error_callback(void *arg)
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
LVRelState *errinfo = arg;
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
switch (errinfo->phase)
{
case VACUUM_ERRCB_PHASE_SCAN_HEAP:
if (BlockNumberIsValid(errinfo->blkno))
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
{
if (OffsetNumberIsValid(errinfo->offnum))
Message style improvements
2021-08-07 12:09:22 +02:00
errcontext("while scanning block %u offset %u of relation \"%s.%s\"",
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
errinfo->blkno, errinfo->offnum, errinfo->relnamespace, errinfo->relname);
else
errcontext("while scanning block %u of relation \"%s.%s\"",
errinfo->blkno, errinfo->relnamespace, errinfo->relname);
}
Improve the vacuum error context phase information. We were displaying the wrong phase information for 'info' message in the index clean up phase because we were switching to the previous phase a bit early. We were also not displaying context information for heap phase unless the block number is valid which is fine for error cases but for messages at 'info' or lower error level it appears to be inconsistent with index phase information. Reported-by: Sawada Masahiko Author: Sawada Masahiko Reviewed-by: Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CA+fd4k4HcbhPnCs7paRTw1K-AHin8y4xKomB9Ru0ATw0UeTy2w@mail.gmail.com
2020-08-24 04:46:19 +02:00
else
errcontext("while scanning relation \"%s.%s\"",
errinfo->relnamespace, errinfo->relname);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
break;
case VACUUM_ERRCB_PHASE_VACUUM_HEAP:
if (BlockNumberIsValid(errinfo->blkno))
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
{
if (OffsetNumberIsValid(errinfo->offnum))
Message style improvements
2021-08-07 12:09:22 +02:00
errcontext("while vacuuming block %u offset %u of relation \"%s.%s\"",
Add additional information in the vacuum error context. The additional information added will be an offset number for heap operations. This information will help us in finding the exact tuple due to which the error has occurred. Author: Mahendra Singh Thalor and Amit Kapila Reviewed-by: Sawada Masahiko, Justin Pryzby and Amit Kapila Discussion: https://postgr.es/m/CAKYtNApK488TDF4bMbw+1QH8HJf9cxdNDXquhU50TK5iv_FtCQ@mail.gmail.com
2020-08-26 06:10:52 +02:00
errinfo->blkno, errinfo->offnum, errinfo->relnamespace, errinfo->relname);
else
errcontext("while vacuuming block %u of relation \"%s.%s\"",
errinfo->blkno, errinfo->relnamespace, errinfo->relname);
}
Improve the vacuum error context phase information. We were displaying the wrong phase information for 'info' message in the index clean up phase because we were switching to the previous phase a bit early. We were also not displaying context information for heap phase unless the block number is valid which is fine for error cases but for messages at 'info' or lower error level it appears to be inconsistent with index phase information. Reported-by: Sawada Masahiko Author: Sawada Masahiko Reviewed-by: Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CA+fd4k4HcbhPnCs7paRTw1K-AHin8y4xKomB9Ru0ATw0UeTy2w@mail.gmail.com
2020-08-24 04:46:19 +02:00
else
errcontext("while vacuuming relation \"%s.%s\"",
errinfo->relnamespace, errinfo->relname);
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
break;
case VACUUM_ERRCB_PHASE_VACUUM_INDEX:
errcontext("while vacuuming index \"%s\" of relation \"%s.%s\"",
errinfo->indname, errinfo->relnamespace, errinfo->relname);
break;
case VACUUM_ERRCB_PHASE_INDEX_CLEANUP:
errcontext("while cleaning up index \"%s\" of relation \"%s.%s\"",
errinfo->indname, errinfo->relnamespace, errinfo->relname);
break;
case VACUUM_ERRCB_PHASE_TRUNCATE:
if (BlockNumberIsValid(errinfo->blkno))
errcontext("while truncating relation \"%s.%s\" to %u blocks",
errinfo->relnamespace, errinfo->relname, errinfo->blkno);
break;
case VACUUM_ERRCB_PHASE_UNKNOWN:
default:
return; /* do nothing; the errinfo may not be
* initialized */
}
}
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
/*
* Updates the information required for vacuum error callback. This also saves
* the current information which can be later restored via restore_vacuum_error_info.
*/
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
static void
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
update_vacuum_error_info(LVRelState *vacrel, LVSavedErrInfo *saved_vacrel,
int phase, BlockNumber blkno, OffsetNumber offnum)
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
if (saved_vacrel)
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
saved_vacrel->offnum = vacrel->offnum;
saved_vacrel->blkno = vacrel->blkno;
saved_vacrel->phase = vacrel->phase;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
}
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->blkno = blkno;
vacrel->offnum = offnum;
vacrel->phase = phase;
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
}
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
/*
* Restores the vacuum information saved via a prior call to update_vacuum_error_info.
*/
static void
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
restore_vacuum_error_info(LVRelState *vacrel,
const LVSavedErrInfo *saved_vacrel)
Improve vacuum error context handling. Use separate functions to save and restore error context information as that made code easier to understand.  Also, make it clear that the index information required for error context is sane. Author: Andres Freund, Justin Pryzby, Amit Kapila Backpatch-through: 13, where it was introduced Discussion: https://postgr.es/m/CAA4eK1LWo+v1OWu=Sky27GTGSCuOmr7iaURNbc5xz6jO+SaPeA@mail.gmail.com
2020-07-01 04:28:36 +02:00
{
Simplify state managed by VACUUM. Reorganize the state struct used by VACUUM -- group related items together to make it easier to understand. Also stop relying on stack variables inside lazy_scan_heap() -- move those into the state struct instead. Doing things this way simplifies large groups of related functions whose function signatures had a lot of unnecessary redundancy. Switch over to using int64 for the struct fields used to count things that are reported to the user via log_autovacuum and VACUUM VERBOSE output. We were using double, but that doesn't seem to have any advantages. Using int64 makes it possible to add assertions that verify that the first pass over the heap (pruning) encounters precisely the same number of LP_DEAD items that get deleted from indexes later on, in the second pass over the heap. These assertions will be added in later commits. Finally, adjust the signatures of functions with IndexBulkDeleteResult pointer arguments in cases where there was ambiguity about whether or not the argument relates to a single index or all indexes. Functions now use the idiom that both ambulkdelete() and amvacuumcleanup() have always used (where appropriate): accept a mutable IndexBulkDeleteResult pointer argument, and return a result IndexBulkDeleteResult pointer to caller. Author: Peter Geoghegan <pg@bowt.ie> Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com> Reviewed-By: Robert Haas <robertmhaas@gmail.com> Discussion: https://postgr.es/m/CAH2-WzkeOSYwC6KNckbhk2b1aNnWum6Yyn0NKP9D-Hq1LGTDPw@mail.gmail.com
2021-04-05 22:21:44 +02:00
vacrel->blkno = saved_vacrel->blkno;
vacrel->offnum = saved_vacrel->offnum;
vacrel->phase = saved_vacrel->phase;
Introduce vacuum errcontext to display additional information. The additional information displayed will be block number for error occurring while processing heap and index name for error occurring while processing the index. This will help us in diagnosing the problems that occur during a vacuum. For ex. due to corruption (either caused by bad hardware or by some bug) if we get some error while vacuuming, it can help us identify the block in heap and or additional index information. It sets up an error context callback to display additional information with the error. During different phases of vacuum (heap scan, heap vacuum, index vacuum, index clean up, heap truncate), we update the error context callback to display appropriate information. We can extend it to a bit more granular level like adding the phases for FSM operations or for prefetching the blocks while truncating. However, I felt that it requires adding many more error callback function calls and can make the code a bit complex, so left those for now. Author: Justin Pryzby, with few changes by Amit Kapila Reviewed-by: Alvaro Herrera, Amit Kapila, Andres Freund, Michael Paquier and Sawada Masahiko Discussion: https://www.postgresql.org/message-id/20191120210600.GC30362@telsasoft.com
2020-03-30 04:03:38 +02:00
}