Teach VACUUM VERBOSE to report on pages deleted by the _current_ VACUUM
operation -- these are newly deleted pages. VACUUM VERBOSE continues to
report on the total number of deleted pages in the entire index (no
change there). The former is a subset of the latter.
The distinction between each category of deleted index page only arises
with index AMs where page deletion is supported and is decoupled from
page recycling for performance reasons.
This is follow-up work to commit e5d8a999, which made nbtree store
64-bit XIDs (not 32-bit XIDs) in pages at the point at which they're
deleted. Note that the btm_last_cleanup_num_delpages metapage field
added by that commit usually gets set to pages_newly_deleted. The
exceptions (the scenarios in which they're not equal) all seem to be
tricky cases for the implementation (of page deletion and recycling) in
general.
Author: Peter Geoghegan <pg@bowt.ie>
Discussion: https://postgr.es/m/CAH2-WznpdHvujGUwYZ8sihX%3Dd5u-tRYhi-F4wnV2uN2zHpMUXw%40mail.gmail.com
Otherwise we risk "leaking" deleted pages by making them non-recyclable
indefinitely. Commit 6655a729 did the same thing for deleted pages in
GiST indexes. That work was used as a starting point here.
Stop storing an XID indicating the oldest bpto.xact across all deleted
though unrecycled pages in nbtree metapages. There is no longer any
reason to care about that condition/the oldest XID. It only ever made
sense when wraparound was something _bt_vacuum_needs_cleanup() had to
consider.
The btm_oldest_btpo_xact metapage field has been repurposed and renamed.
It is now btm_last_cleanup_num_delpages, which is used to remember how
many non-recycled deleted pages remain from the last VACUUM (in practice
its value is usually the precise number of pages that were _newly
deleted_ during the specific VACUUM operation that last set the field).
The general idea behind storing btm_last_cleanup_num_delpages is to use
it to give _some_ consideration to non-recycled deleted pages inside
_bt_vacuum_needs_cleanup() -- though never too much. We only really
need to avoid leaving a truly excessive number of deleted pages in an
unrecycled state forever. We only do this to cover certain narrow cases
where no other factor makes VACUUM do a full scan, and yet the index
continues to grow (and so actually misses out on recycling existing
deleted pages).
These metapage changes result in a clear user-visible benefit: We no
longer trigger full index scans during VACUUM operations solely due to
the presence of only 1 or 2 known deleted (though unrecycled) blocks
from a very large index. All that matters now is keeping the costs and
benefits in balance over time.
Fix an issue that has been around since commit 857f9c36, which added the
"skip full scan of index" mechanism (i.e. the _bt_vacuum_needs_cleanup()
logic). The accuracy of btm_last_cleanup_num_heap_tuples accidentally
hinged upon _when_ the source value gets stored. We now always store
btm_last_cleanup_num_heap_tuples in btvacuumcleanup(). This fixes the
issue because IndexVacuumInfo.num_heap_tuples (the source field) is
expected to accurately indicate the state of the table _after_ the
VACUUM completes inside btvacuumcleanup().
A backpatchable fix cannot easily be extracted from this commit. A
targeted fix for the issue will follow in a later commit, though that
won't happen today.
I (pgeoghegan) have chosen to remove any mention of deleted pages in the
documentation of the vacuum_cleanup_index_scale_factor GUC/param, since
the presence of deleted (though unrecycled) pages is no longer of much
concern to users. The vacuum_cleanup_index_scale_factor description in
the docs now seems rather unclear in any case, and it should probably be
rewritten in the near future. Perhaps some passing mention of page
deletion will be added back at the same time.
Bump XLOG_PAGE_MAGIC due to nbtree WAL records using full XIDs now.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Masahiko Sawada <sawada.mshk@gmail.com>
Discussion: https://postgr.es/m/CAH2-WznpdHvujGUwYZ8sihX=d5u-tRYhi-F4wnV2uN2zHpMUXw@mail.gmail.com
Consolidate discussion of how VACUUM places pages in the FSM for
recycling by adding a new section that comes after discussion of page
deletion. This structure reflects the fact that page recycling is
explicitly decoupled from page deletion in Lanin & Shasha's paper. Page
recycling in nbtree is an implementation of what the paper calls "the
drain technique".
This decoupling is an important concept for nbtree VACUUM. Searchers
have to detect and recover from concurrent page deletions, but they will
never have to reason about concurrent page recycling. Recycling can
almost always be thought of as a low level garbage collection operation
that asynchronously frees the physical space that backs a logical tree
node. Almost all code need only concern itself with logical tree nodes.
(Note that "logical tree node" is not currently a term of art in the
nbtree code -- this all works implicitly.)
This is preparation for an upcoming patch that teaches nbtree VACUUM to
remember the details of pages that it deletes on the fly, in local
memory. This enables the same VACUUM operation to consider placing its
own deleted pages in the FSM later on, when it reaches the end of
btvacuumscan().
Discuss VACUUM's linear scan after discussion of tuple deletion by
VACUUM, but before discussion of page deletion by VACUUM. This
progression is a lot more natural.
Also tweak the wording a little. It seems unnecessary to talk about how
it worked prior to PostgreSQL 8.2.
Add some additional defensive checks in the second phase of index
deletion to detect and report index corruption during VACUUM, and to
avoid having VACUUM become stuck in more cases. The code is still not
robust in the presence of a circular chain of sibling links, though it's
not clear whether that really matters. This is follow-up work to commit
3a01f68e.
The new defensive checks rely on the assumption that there can be no
more than one VACUUM operation running for an index at any given time.
Remove an old comment suggesting that multiple concurrent VACUUMs need
to be considered here. This concern now seems highly unlikely to have
any real validity, since we clearly rely on the same assumption in
several other places. For example, there are much more recent comments
that appear in the same function (added by commit efada2b8e9) that make
the same assumption.
Also add a CHECK_FOR_INTERRUPTS() to the relevant code path. Contrary
to comments added by commit 3a01f68e, it is actually possible to handle
interrupts here, at least in the common case where processing takes
place at the leaf level. We only hold a pin on leafbuf/target page when
stepping right at the leaf level.
No backpatch due to the lack of complaints following hardening added to
the same area by commit 3a01f68e.
The prototype calls the second argument of
pgstat_progress_update_multi_param() "index", and some callers name
their local variable that way. But when the surrounding code deals
with index relations, this is confusing, and in at least one case
shadowed another variable that is referring to an index relation.
Adjust those call sites to have clearer local variable naming, similar
to existing callers in indexcmds.c.
The latestRemovedXid values used by nbtree deletion operations are
determined by _bt_delitems_delete()'s caller, so there is no reason to
pass a separate heapRel argument.
Oversight in commit d168b66682.
Teach nbtree and heapam to cooperate in order to eagerly remove
duplicate tuples representing dead MVCC versions. This is "bottom-up
deletion". Each bottom-up deletion pass is triggered lazily in response
to a flood of versions on an nbtree leaf page. This usually involves a
"logically unchanged index" hint (these are produced by the executor
mechanism added by commit 9dc718bd).
The immediate goal of bottom-up index deletion is to avoid "unnecessary"
page splits caused entirely by version duplicates. It naturally has an
even more useful effect, though: it acts as a backstop against
accumulating an excessive number of index tuple versions for any given
_logical row_. Bottom-up index deletion complements what we might now
call "top-down index deletion": index vacuuming performed by VACUUM.
Bottom-up index deletion responds to the immediate local needs of
queries, while leaving it up to autovacuum to perform infrequent clean
sweeps of the index. The overall effect is to avoid certain
pathological performance issues related to "version churn" from UPDATEs.
The previous tableam interface used by index AMs to perform tuple
deletion (the table_compute_xid_horizon_for_tuples() function) has been
replaced with a new interface that supports certain new requirements.
Many (perhaps all) of the capabilities added to nbtree by this commit
could also be extended to other index AMs. That is left as work for a
later commit.
Extend deletion of LP_DEAD-marked index tuples in nbtree by adding logic
to consider extra index tuples (that are not LP_DEAD-marked) for
deletion in passing. This increases the number of index tuples deleted
significantly in many cases. The LP_DEAD deletion process (which is now
called "simple deletion" to clearly distinguish it from bottom-up
deletion) won't usually need to visit any extra table blocks to check
these extra tuples. We have to visit the same table blocks anyway to
generate a latestRemovedXid value (at least in the common case where the
index deletion operation's WAL record needs such a value).
Testing has shown that the "extra tuples" simple deletion enhancement
increases the number of index tuples deleted with almost any workload
that has LP_DEAD bits set in leaf pages. That is, it almost never fails
to delete at least a few extra index tuples. It helps most of all in
cases that happen to naturally have a lot of delete-safe tuples. It's
not uncommon for an individual deletion operation to end up deleting an
order of magnitude more index tuples compared to the old naive approach
(e.g., custom instrumentation of the patch shows that this happens
fairly often when the regression tests are run).
Add a further enhancement that augments simple deletion and bottom-up
deletion in indexes that make use of deduplication: Teach nbtree's
_bt_delitems_delete() function to support granular TID deletion in
posting list tuples. It is now possible to delete individual TIDs from
posting list tuples provided the TIDs have a tableam block number of a
table block that gets visited as part of the deletion process (visiting
the table block can be triggered directly or indirectly). Setting the
LP_DEAD bit of a posting list tuple is still an all-or-nothing thing,
but that matters much less now that deletion only needs to start out
with the right _general_ idea about which index tuples are deletable.
Bump XLOG_PAGE_MAGIC because xl_btree_delete changed.
No bump in BTREE_VERSION, since there are no changes to the on-disk
representation of nbtree indexes. Indexes built on PostgreSQL 12 or
PostgreSQL 13 will automatically benefit from bottom-up index deletion
(i.e. no reindexing required) following a pg_upgrade. The enhancement
to simple deletion is available with all B-Tree indexes following a
pg_upgrade, no matter what PostgreSQL version the user upgrades from.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Heikki Linnakangas <hlinnaka@iki.fi>
Reviewed-By: Victor Yegorov <vyegorov@gmail.com>
Discussion: https://postgr.es/m/CAH2-Wzm+maE3apHB8NOtmM=p-DO65j2V5GzAWCOEEuy3JZgb2g@mail.gmail.com
Add an executor aminsert() hint mechanism that informs index AMs that
the incoming index tuple (the tuple that accompanies the hint) is not
being inserted by execution of an SQL statement that logically modifies
any of the index's key columns.
The hint is received by indexes when an UPDATE takes place that does not
apply an optimization like heapam's HOT (though only for indexes where
all key columns are logically unchanged). Any index tuple that receives
the hint on insert is expected to be a duplicate of at least one
existing older version that is needed for the same logical row. Related
versions will typically be stored on the same index page, at least
within index AMs that apply the hint.
Recognizing the difference between MVCC version churn duplicates and
true logical row duplicates at the index AM level can help with cleanup
of garbage index tuples. Cleanup can intelligently target tuples that
are likely to be garbage, without wasting too many cycles on less
promising tuples/pages (index pages with little or no version churn).
This is infrastructure for an upcoming commit that will teach nbtree to
perform bottom-up index deletion. No index AM actually applies the hint
just yet.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Victor Yegorov <vyegorov@gmail.com>
Discussion: https://postgr.es/m/CAH2-Wz=CEKFa74EScx_hFVshCOn6AA5T-ajFASTdzipdkLTNQQ@mail.gmail.com
"Ordering stuff" refered to a _bt_first() call to _bt_orderkeys().
However, the _bt_orderkeys() function was renamed to
_bt_preprocess_keys() by commit fa5c8a055a.
_bt_preprocess_keys() is directly referenced just after the removed
comment already, which seems sufficient.
User-visible log messages should go through ereport(), so they are
subject to translation. Many remaining elog(LOG) calls are really
debugging calls.
Reviewed-by: Alvaro Herrera <alvherre@alvh.no-ip.org>
Reviewed-by: Michael Paquier <michael@paquier.xyz>
Reviewed-by: Noah Misch <noah@leadboat.com>
Discussion: https://www.postgresql.org/message-id/flat/92d6f545-5102-65d8-3c87-489f71ea0a37%40enterprisedb.com
Streamline handling of the various strategies that we have to avoid a
page split in nbtinsert.c. When it looks like a leaf page is about to
overflow, we now perform deleting LP_DEAD items and deduplication in one
central place. This greatly simplifies _bt_findinsertloc().
This has an independently useful consequence: nbtree no longer relies on
the BTP_HAS_GARBAGE page level flag/hint for anything important. We
still set and unset the flag in the same way as before, but it's no
longer treated as a gating condition when considering if we should check
for already-set LP_DEAD bits. This happens at the point where the page
looks like it might have to be split anyway, so simply checking the
LP_DEAD bits in passing is practically free. This avoids missing
LP_DEAD bits just because the page-level hint is unset, which is
probably reasonably common (e.g. it happens when VACUUM unsets the
page-level flag without actually removing index tuples whose LP_DEAD-bit
was set recently, after the VACUUM operation began but before it reached
the leaf page in question).
Note that this isn't a big behavioral change compared to PostgreSQL 13.
We were already checking for set LP_DEAD bits regardless of whether the
BTP_HAS_GARBAGE page level flag was set before we considered doing a
deduplication pass. This commit only goes slightly further by doing the
same check for all indexes, even indexes where deduplication won't be
performed.
We don't completely remove the BTP_HAS_GARBAGE flag. We still rely on
it as a gating condition with pg_upgrade'd indexes from before B-tree
version 4/PostgreSQL 12. That makes sense because we sometimes have to
make a choice among pages full of duplicates when inserting a tuple with
pre version 4 indexes. It probably still pays to avoid accessing the
line pointer array of a page there, since it won't yet be clear whether
we'll insert on to the page in question at all, let alone split it as a
result.
Author: Peter Geoghegan <pg@bowt.ie>
Reviewed-By: Victor Yegorov <vyegorov@gmail.com>
Discussion: https://postgr.es/m/CAH2-Wz%3DYpc1PDdk8OVJDChGJBjT06%3DA0Mbv9HyTLCsOknGcUFg%40mail.gmail.com
Stop naming special area/opaque pointer variables 'lpageop' in contexts
where it doesn't make sense. This is a holdover from a time when logic
that performs tasks that are now spread across _bt_insertonpg(),
_bt_findinsertloc(), and _bt_split() was more centralized. 'lpageop'
denotes "left page", which doesn't make sense outside of contexts in
which there isn't also a right page.
Also acquire page flag variables up front within _bt_insertonpg(). This
makes it closer to _bt_split() following refactoring commit bc3087b626.
This allows the page split and retail insert paths to both make use of
the same variables.
Only a basic logic bug in a _bt_insertonpg() caller could lead to a
violation of this invariant (index corruption won't do it). A "can't
happen" error seems inappropriate (it is arbitrary at best).
Demote the error to a simple assertion. This matches similar nearby
sanity checks.
The description of how LP_DEAD bit setting by index scans works
following commit 2ed5b87f was rather unclear. Clean that up a bit.
Also refer to LP_DEAD bit setting within _bt_check_unique() at the start
of the same section. This mechanism may actually be more important than
the generic kill_prior_tuple mechanism that the section focuses on, so
it at least deserves to be mentioned in passing.
wal_consistency_checking indicated an inconsistency in certain cases
involving nbtree page deletion. The underlying issue is that there was
a minor difference between the page image produced after a REDO routine
ran and the corresponding page image following original execution.
This harmless inconsistency has been around forever. We more or less
expect total consistency among even deleted nbtree pages these days,
though, so this won't do anymore.
To fix, tweak the REDO routine to match original execution.
Oversight in commit f47b5e13.
Logic for counting heap TIDs from posting list tuples (added by commit
0d861bbb) was faulty. It didn't count any TIDs/index tuples in the
event of no callback being set. This meant that we incorrectly counted
no index tuples in clean-up only VACUUMs, which could lead to
pg_class.reltuples being spuriously set to 0 in affected indexes.
To fix, go back to counting items from the page in cases where there is
no callback. This approach isn't very accurate, but it works well
enough in practice while avoiding the expense of accessing every index
tuple during cleanup-only VACUUMs.
Author: Peter Geoghegan <pg@bowt.ie>
Reported-By: Jehan-Guillaume de Rorthais <jgdr@dalibo.com>
https://postgr.es/m/20201023174451.69e358f1@firost
Backpatch: 13-, where nbtree deduplication was introduced
Certain background workers initiate parallel queries while
debug_query_string==NULL, at which point they attempted strlen(NULL) and
died to SIGSEGV. Older debug_query_string observers allow NULL, so do
likewise in these newer ones. Back-patch to v11, where commit
7de4a1bcc5 introduced the first of these.
Discussion: https://postgr.es/m/20201014022636.GA1962668@rfd.leadboat.com
For parallel btree scan to work for array of scan keys, it should reach
BTPARALLEL_DONE state once for every distinct combination of array keys.
This is required to ensure that the parallel workers don't try to seize
blocks at the same time for different scan keys. We missed to update this
state when we discovered that the scan keys can't be satisfied.
Author: James Hunter
Reviewed-by: Amit Kapila
Tested-by: Justin Pryzby
Backpatch-through: 10, where it was introduced
Discussion: https://postgr.es/m/4248CABC-25E3-4809-B4D0-128E1BAABC3C@amazon.com
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
Now that xmin isn't needed for GetSnapshotData() anymore, it leads to
unnecessary cacheline ping-pong to have it in PGXACT, as it is updated
considerably more frequently than the other PGXACT members.
After the changes in dc7420c2c9, this is a very straight-forward change.
For highly concurrent, snapshot acquisition heavy, workloads this change alone
can significantly increase scalability. E.g. plain pgbench on a smaller 2
socket machine gains 1.07x for read-only pgbench, 1.22x for read-only pgbench
when submitting queries in batches of 100, and 2.85x for batches of 100
'SELECT';. The latter numbers are obviously not to be expected in the
real-world, but micro-benchmark the snapshot computation
scalability (previously spending ~80% of the time in GetSnapshotData()).
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
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
Make the nbtree page split REDO routine consistent with original
execution in its approach to acquiring and releasing buffer locks (at
least for pages on the tree level of the page being split). This brings
btree_xlog_split() in line with btree_xlog_unlink_page(), which was
taught to couple buffer locks by commit 9a9db08a.
Note that the precise order in which we both acquire and release sibling
buffer locks in btree_xlog_split() now matches original execution
exactly (the precise order in which the locks are released probably
doesn't matter much, but we might as well be consistent about it).
The rule for nbtree REDO routines from here on is that same-level locks
should be acquired in an order that's consistent with original
execution. It's not practical to have a similar rule for cross-level
page locks, since for the most part original execution holds those locks
for a period that spans multiple atomic actions/WAL records. It's also
not necessary, because clearly the cross-level lock coupling is only
truly needed during original execution because of the presence of
concurrent inserters.
This is not a bug fix (unlike the similar aforementioned commit, commit
9a9db08a). The immediate reason to tighten things up in this area is to
enable an upcoming enhancement to contrib/amcheck that allows it to
verify that sibling links are in agreement with only an AccessShareLock
(this check produced false positives when run on a replica server on
account of the inconsistency fixed by this commit). But that's not the
only reason to be stricter here.
It is generally useful to make locking on replicas be as close to what
happens during original execution as practically possible. It makes it
less likely that hard to catch bugs will slip in in the future. The
previous state of affairs seems to be a holdover from before the
introduction of Hot Standby, when buffer lock acquisitions during
recovery were totally unnecessary. See also: commit 3bbf668d, which
tightened things up in this area a few years after the introduction of
Hot Standby.
Discussion: https://postgr.es/m/CAH2-Wz=465cJj11YXD9RKH8z=nhQa2dofOZ_23h67EXUGOJ00Q@mail.gmail.com
Make the nbtree page split REDO routine variable names consistent with
_bt_split() (which handles the original execution of page splits).
These names make the code easier to follow by making the distinction
between the original page and the left half of the split clear. (The
left half of the split page is a temp page that REDO creates to replace
the origpage contents.)
Also reduce the elevel used when adding a new high key to the temp page
from PANIC to ERROR to be consistent. We already only raise an ERROR
when data item PageAddItem() temp page calls fail.
Currently, page unlink leaves remaining items "as is", but replay of
corresponding WAL-record re-initializes page leaving it with no items.
For the sake of consistency, this commit makes primary delete all the items
during page unlink as well.
Thanks to this change, we now don't mask contents of deleted btree page for
WAL consistency checking.
Discussion: https://postgr.es/m/CAPpHfdt_OTyQpXaPJcWzV2N-LNeNJseNB-K_A66qG%3DL518VTFw%40mail.gmail.com
Author: Alexander Korotkov
Reviewed-by: Peter Geoghegan
It was possible for the logic used by backward scans (which must reason
about concurrent page splits/deletions in its own peculiar way) to
become confused when running on a replica. Concurrent replay of a WAL
record that describes the second phase of page deletion could cause
_bt_walk_left() to get confused. btree_xlog_unlink_page() simply failed
to adhere to the same locking protocol that we use on the primary, which
is obviously wrong once you consider these two disparate functions
together. This bug is present in all stable branches.
More concretely, the problem was that nothing stopped _bt_walk_left()
from observing inconsistencies between the deletion's target page and
its original sibling pages when running on a replica. This is true even
though the second phase of page deletion is supposed to work as a single
atomic action. Queries running on replicas raised "could not find left
sibling of block %u in index %s" can't-happen errors when they went back
to their scan's "original" page and observed that the page has not been
marked deleted (even though it really was concurrently deleted).
There is no evidence that this actually happened in the real world. The
issue came to light during unrelated feature development work. Note
that _bt_walk_left() is the only code that cares about the difference
between a half-dead page and a fully deleted page that isn't also
exclusively used by nbtree VACUUM (unless you include contrib/amcheck
code). It seems very likely that backward scans are the only thing that
could become confused by the inconsistency. Even amcheck's complex
bt_right_page_check_scankey() dance was unaffected.
To fix, teach btree_xlog_unlink_page() to lock the left sibling, target,
and right sibling pages in that order before releasing any locks (just
like _bt_unlink_halfdead_page()). This is the simplest possible
approach. There doesn't seem to be any opportunity to be more clever
about lock acquisition in the REDO routine, and it hardly seems worth
the trouble in any case.
This fix might enable contrib/amcheck verification of leaf page sibling
links with only an AccessShareLock on the relation. An amcheck patch
from Andrey Borodin was rejected back in January because it clashed with
btree_xlog_unlink_page()'s lax approach to locking pages. It now seems
likely that the real problem was with btree_xlog_unlink_page(), not the
patch.
This is a low severity, low likelihood bug, so no backpatch.
Author: Michail Nikolaev
Diagnosed-By: Michail Nikolaev
Discussion: https://postgr.es/m/CANtu0ohkR-evAWbpzJu54V8eCOtqjJyYp3PQ_SGoBTRGXWhWRw@mail.gmail.com
Add a documenting assertion that's similar to the nearby assertion added
by commit cd8c73a3. This conveys that the entire call to _bt_pagedel()
does no work if it isn't possible to get a descent stack for the initial
scanblkno page.
This allows AM-specific knowledge to be applied during creation of
pg_amop and pg_amproc entries. Specifically, the AM knows better than
core code which entries to consider as required or optional. Giving
the latter entries the appropriate sort of dependency allows them to
be dropped without taking out the whole opclass or opfamily; which
is something we'd like to have to correct obsolescent entries in
extensions.
This callback also opens the door to performing AM-specific validity
checks during opclass creation, rather than hoping than an opclass
developer will remember to test with "amvalidate". For the most part
I've not actually added any such checks yet; that can happen in a
follow-on patch. (Note that we shouldn't remove any tests from
"amvalidate", as those are still needed to cross-check manually
constructed entries in the initdb data. So adding tests to
"amadjustmembers" will be somewhat duplicative, but it seems like
a good idea anyway.)
Patch by me, reviewed by Alexander Korotkov, Hamid Akhtar, and
Anastasia Lubennikova.
Discussion: https://postgr.es/m/4578.1565195302@sss.pgh.pa.us
Holding just a buffer pin (with no buffer lock) on an nbtree buffer/page
provides very weak guarantees, especially compared to heapam, where it's
often safe to read a page while only holding a buffer pin. This commit
has Valgrind enforce the following rule: it is never okay to access an
nbtree buffer without holding both a pin and a lock on the buffer.
A draft version of this patch detected questionable code that was
cleaned up by commits fa7ff642 and 7154aa16. The code in question used
to access an nbtree buffer page's special/opaque area with no buffer
lock (only a buffer pin). This practice (which isn't obviously unsafe)
is hereby formally disallowed in nbtree. There doesn't seem to be any
reason to allow it, and banning it keeps things simple for Valgrind.
The new checks are implemented by adding custom nbtree client requests
(located in LockBuffer() wrapper functions); these requests are
"superimposed" on top of the generic bufmgr.c Valgrind client requests
added by commit 1e0dfd16. No custom resource management cleanup code is
needed to undo the effects of marking buffers as non-accessible under
this scheme.
Author: Peter Geoghegan
Reviewed-By: Anastasia Lubennikova, Georgios Kokolatos
Discussion: https://postgr.es/m/CAH2-WzkLgyN3zBvRZ1pkNJThC=xi_0gpWRUb_45eexLH1+k2_Q@mail.gmail.com
There is no advantage to attempting deduplication for a unique index
during CREATE INDEX, since there cannot possibly be any duplicates.
Doing so wastes cycles due to unnecessary copying. Make sure that we
avoid it consistently.
We already avoided unique index deduplication in the case where there
were some spool2 tuples to merge. That didn't account for the fact that
spool2 is removed early/unset in the common case where it has no tuples
that need to be merged (i.e. it failed to account for the "spool2 turns
out to be unnecessary" optimization in _bt_spools_heapscan()).
Oversight in commit 0d861bbb, which added nbtree deduplication
Backpatch: 13-, where nbtree deduplication was introduced.
nbtree index builds cannot write out an empty page. That would mean
that there was no way to create a pivot tuple pointing to the page one
level up, since _bt_truncate() generates one based on page's firstright
tuple.
Replace the unnecessary PageIsEmpty() check with an assertion that
checks that the page has space for at least two line pointers (the
would-be high key line pointer, plus at least one valid "data item"
tuple line pointer).
The PageIsEmpty() check was added by commit 5d9f146c over 20 years ago.
It looks like it has always been unnecessary.
Make some of the variable names in _bt_search() consistent with
corresponding variables within _bt_getstackbuf(). This naming scheme is
clearer because the variable names always express a relationship between
the currently locked buffer/page and some other page.
Commit 0d861bbb, which added deduplication to nbtree, had
_bt_check_unique() pass a TID to table_index_fetch_tuple_check() that
isn't safe to mutate. table_index_fetch_tuple_check()'s tid argument is
modified when the TID in question is not the latest visible tuple in a
hot chain, though this wasn't documented.
To fix, go back to using a local copy of the TID in _bt_check_unique(),
and update comments above table_index_fetch_tuple_check().
Backpatch: 13-, where B-Tree deduplication was introduced.
It was possible for deduplication's single value strategy to mistakenly
believe that a very small duplicate tuple counts as one of the six large
tuples that it aims to leave behind after the page finally splits. This
could cause slightly suboptimal space utilization with very low
cardinality indexes, though only under fairly narrow conditions.
To fix, be particular about what kind of tuple counts as a
maxpostingsize-capped tuple. This avoids confusion in the event of a
small tuple that gets "wedged" between two large tuples, where all
tuples on the page are duplicates of the same value.
Discussion: https://postgr.es/m/CAH2-Wz=Y+sgSFc-O3LpiZX-POx2bC+okec2KafERHuzdVa7-rQ@mail.gmail.com
Backpatch: 13-, where deduplication was introduced (by commit 0d861bbb)
Commit 624686abcf added an assertion that verified that _bt_search
successfully relocated the leaf page undergoing deletion. Page deletion
cannot deal with the case where the descent stack is to the right of the
page, so this seemed critical (deletion can only handle the case where
the descent stack is to the left of the leaf/target page). However, the
assertion went a bit too far.
Since only a buffer pin is held on the leaf page throughout the call to
_bt_search, nothing guarantees that it can't have split during this
small window. And if does actually split, _bt_search may end up
"relocating" a page to the right of the original target leaf page. This
scenario seems extremely unlikely, but it must still be considered.
Remove the assertion, and document how we cope in this scenario.
_bt_killitems marks btree items dead when a scan leaves the page where
they live, but it does so with only share lock (to improve concurrency).
This was historicall okay, since killing a dead item has no
consequences. However, with the advent of data checksums and
wal_log_hints, this action incurs a WAL full-page-image record of the
page. Multiple concurrent processes would write the same page several
times, leading to WAL bloat. The probability of this happening can be
reduced by only killing items if they're not already dead, so change the
code to do that.
The problem could eliminated completely by having _bt_killitems upgrade
to exclusive lock upon seeing a killable item, but that would reduce
concurrency so it's considered a cure worse than the disease.
Backpatch all the way back to 9.5, since wal_log_hints was introduced in
9.4.
Author: Masahiko Sawada <masahiko.sawada@2ndquadrant.com>
Discussion: https://postgr.es/m/CA+fd4k6PeRj2CkzapWNrERkja5G0-6D-YQiKfbukJV+qZGFZ_Q@mail.gmail.com
Includes some manual cleanup of places that pgindent messed up,
most of which weren't per project style anyway.
Notably, it seems some people didn't absorb the style rules of
commit c9d297751, because there were a bunch of new occurrences
of function calls with a newline just after the left paren, all
with faulty expectations about how the rest of the call would get
indented.
Restructure the function that locates the root of the to-be-deleted
subtree during nbtree page deletion. Handle the conditions that make
page deletion unsafe in a slightly more uniform way, and acknowledge the
fact that the behavior with incomplete splits on internal pages is
different (as pointed out in the nbtree README as of commit 35bc0ec7).
Also invent new terminology that avoids ambiguity around which pages are
about to be deleted. Consistently use the term "to-be-deleted subtree",
not the ambiguous term "branch".
We were calling the subtree parent page the "top parent page", but that
was quite misleading. The top parent page usually refers to a page
unlinked from its siblings and marked deleted (during the second stage
of page deletion). There was one kind of top parent page that we merely
removed a downlink from, and another kind of top parent page that we
actually marked deleted. Eliminate the ambiguity by inventing a new
term ("subtree parent page") that refers to the former kind of page
only.
Factor out code common to _bt_lock_branch_parent() and _bt_pagedel()
into a new utility function. This new function is used to check that
the left sibling of a deletion target page does not have the
INCOMPLETE_SPLIT page flag set. If it is set then deletion is unsafe;
there won't be a usable pivot tuple (with a downlink) in the parent page
that points to the deletion target page. The page deletion algorithm is
not prepared to deal with that. Also restructure an existing, related
utility function that checks if the right sibling of the target page has
the ISHALFDEAD page flag set.
This organization highlights the symmetry between the two cases. The
goal is to make the design of page deletion clearer. Both functions
involve a sibling page with a flag that indicates that there was an
interrupted operation (a page split or a page deletion) that resulted in
a page pointed to by sibling pages, but not pointed to in the parent.
And, both functions indicate if page deletion is unsafe due to the
absence of a particular downlink in the parent page.
Remove one of the arguments to btvacuumpage(), and give up on the idea
that it's a recursive function. We now use the term "backtracking" to
refer to the case where an earlier block must be visited to make sure no
tuples that need to be removed were missed.
Advertising btvacuumpage() as a recursive function was unhelpful. In
reality the function always simulates recursion with a loop (it doesn't
actually call itself). This wasn't just necessary as a precaution (per
the comments mentioning tail recursion), though. There is no reliable
natural limit on the number of times we can backtrack.
There are important behavioral difference when "recursing"/backtracking,
mostly related to page deletion. We don't perform page deletion when
backtracking due to the extra complexity. And when we recurse, we're
not performing a physical order scan anymore, so we expect fairly
different conditions to hold for the page. Structuring the code like
this makes it clearer how _bt_pagedel() cooperates with btvacuumpage()
and btvacuumscan() (as established in commit b0229f26 and commit
73a076b0).
Author: Peter Geoghegan
Reviewed-By: Masahiko Sawada
Discussion: https://postgr.es/m/CAH2-WzmRGMDWiLMcb+zagG9652PboNN4Gfcq1Gc_wJL6A716MA@mail.gmail.com
Peter Geoghegan
Thomas Munro
Peter Eisentraut
Bruce Momjian
Noah Misch
Amit Kapila
Tom Lane
Andres Freund
Alexander Korotkov
Alvaro Herrera