This provides a speedup of about 4X when NBuffers is large enough.
There is also a useful reduction in sinval traffic, since we
only do CacheInvalidateSmgr() once not once per fork.
Simon Riggs, reviewed and somewhat revised by Tom Lane
WALSender now woken up after each background flush by WALwriter, avoiding
multi-second replication delay for an all-async commit workload.
Replication delay reduced from 7s with default settings to 200ms and often
much less, allowing significantly reduced data loss at failover.
Andres Freund and Simon Riggs
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some related logic so that we rely on the possibly-outdated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.
When we allowed read-only transactions to skip assigning XIDs
we introduced the possibility that a fully deleted btree page
could be reused. This broke the index link sequence which could
then lead to indexscans silently returning fewer rows than would
have been correct. The actual incidence of silent errors from
this is thought to be very low because of the exact workload
required and locking pre-conditions. Fix is to remove pages only
if index page opaque->btpo.xact precedes RecentGlobalXmin.
Noah Misch, reviewed by Simon Riggs
We used to mimic the way a stack is constructed when descending the tree
during normal GiST inserts, but that was quite complicated during a buffered
build. It was also wrong: in GiST, the left-to-right relationships on
different levels might not match each other, so that when you know the
parent of a child page, you won't necessarily find the parent of the page to
the right of the child page by following the rightlinks at the parent level.
This sometimes led to "could not re-find parent" errors while building a
GiST index.
We now use a simple hash table to track the parent of every internal page.
Whenever a page is split, and downlinks are moved from one page to another,
we update the hash table accordingly. This is also better for performance
than the old method, as we never need to move right to re-find the parent
page, which could take a significant amount of time for buffers that were
created much earlier in the index build.
There were two bugs here: We forgot to call gistFreeBuildBuffers() function
at the end of build, and we passed interXact == true to BufFileCreateTemp,
so the file wasn't automatically cleaned up at end-of-transaction either.
The result of (maintenance_work_mem * 1024) / BLCKSZ doesn't fit in a signed
32-bit integer, if maintenance_work_mem >= 2GB. Use double instead. And
while we're at it, write the calculations in an easier to understand form,
with the intermediary steps written out and commented.
AbortOutOfAnyTransaction failed to do anything if the state it saw on
entry corresponded to failing partway through StartTransaction. I fixed
AbortCurrentTransaction to cope with that case way back in commit
60b2444cc3, but evidently overlooked that
AbortOutOfAnyTransaction should do likewise.
Back-patch to all supported branches. It's not clear that this omission
has any more-than-cosmetic consequences, but it's also not clear that it
doesn't, so back-patching seems the least risky choice.
The only interesting-for-performance case wherein we force heapscan here
is when we're rebuilding the relcache init file, and the only such case
that is likely to be examining a catalog big enough to be syncscanned is
RelationBuildTupleDesc. But the early-exit optimization in that code gets
broken if we start the scan at a random place within the catalog, so that
allowing syncscan is actually a big deoptimization if pg_attribute is large
(at least for the normal case where the rows for core system catalogs have
never been changed since initdb). Hence, prevent syncscan here. Per my
testing pursuant to complaints from Jeff Frost and Greg Sabino Mullane,
though neither of them seem to have actually hit this specific problem.
Back-patch to 8.3, where syncscan was introduced.
If a seqscan encounters many consecutive pages containing only dead tuples,
it can remain in the loop in heapgettup for a long time, and there was no
CHECK_FOR_INTERRUPTS anywhere in that loop. This meant there were
real-world situations where a query would be effectively uncancelable for
long stretches. Add a check placed to occur once per page, which should be
enough to provide reasonable response time without adding any measurable
overhead.
Report and patch by Merlin Moncure (though I tweaked it a bit).
Back-patch to all supported branches.
When we create a temporary copy of the old node buffer, in stack, we mustn't
leak that into any of the long-lived data structures. Before this patch,
when we called gistPopItupFromNodeBuffer(), it got added to the array of
"loaded buffers". After gistRelocateBuildBuffersOnSplit() exits, the
pointer added to the loaded buffers array points to garbage. Often that goes
unnotied, because when we go through the array of loaded buffers to unload
them, buffers with a NULL pageBuffer are ignored, which can often happen by
accident even if the pointer points to garbage.
This patch fixes that by marking the temporary copy in stack explicitly as
temporary, and refrain from adding buffers marked as temporary to the array
of loaded buffers.
While we're at it, initialize nodeBuffer->pageBlocknum to InvalidBlockNumber
and improve comments a bit. This isn't strictly necessary, but makes
debugging easier.
If the amount of freespace on page was less than the amount reserved by
fillfactor, the calculation would underflow.
This fixes bug #6643 reported by Tomonari Katsumata.
When the "hot" members of PGPROC were split off to separate PGXACT structs,
many PGPROC fields referred to in comments were moved to PGXACT, but the
comments were neglected in the commit. Mostly this is just a search/replace
of PGPROC with PGXACT, but the way the dummy PGPROC entries are created for
prepared transactions changed more, making some of the comments totally
bogus.
Noah Misch
When inserting the downlinks for a split gist page, we used hold the locks
on the child pages until the insertion into the parent - and recursively its
parent if it had to be split too - were all completed. Change that so that
the locks on child pages are released after the insertion in the immediate
parent is done, before recursing further up the tree.
This reduces the number of lwlocks that are held simultaneously. Holding
many locks is bad for concurrency, and in extreme cases you can even hit
the limit of 100 simultaneously held lwlocks in a backend. If you're really
unlucky, you can hit the limit while in a critical section, which brings
down the whole system.
This fixes bug #6629 reported by Tom Forbes. Backpatch to 9.1. The page
splitting code was rewritten in 9.1, and the old code did not have this
problem.
Users of asynchronous-commit mode expect there to be a guaranteed maximum
delay before an async commit's WAL records get flushed to disk. The
original version of the walwriter hibernation patch broke that. Add an
extra shared-memory flag to allow async commits to kick the walwriter out
of hibernation mode, without adding any noticeable overhead in cases where
no action is needed.
This patch modifies the walwriter process so that, when it has not found
anything useful to do for many consecutive wakeup cycles, it extends its
sleep time to reduce the server's idle power consumption. It reverts to
normal as soon as it's done any successful flushes. It's still true that
during any async commit, backends check for completed, unflushed pages of
WAL and signal the walwriter if there are any; so that in practice the
walwriter can get awakened and returned to normal operation sooner than the
sleep time might suggest.
Also, improve the checkpointer so that it uses a latch and a computed delay
time to not wake up at all except when it has something to do, replacing a
previous hardcoded 0.5 sec wakeup cycle. This also is primarily useful for
reducing the server's power consumption when idle.
In passing, get rid of the dedicated latch for signaling the walwriter in
favor of using its procLatch, since that comports better with possible
generic signal handlers using that latch. Also, fix a pre-existing bug
with failure to save/restore errno in walwriter's signal handlers.
Peter Geoghegan, somewhat simplified by Tom
At the time we check whether the tuple is dead to all running
transactions, we've already verified that it isn't visible to our
scan, setting hint bits if appropriate. So there's no need to
recheck CLOG for the all-dead test we do just a moment later.
So, add HeapTupleIsSurelyDead() to test the appropriate condition
under the assumption that all relevant hit bits are already set.
Review by Tom Lane.
This patch adjusts the core statistics views to match the decision already
taken for pg_stat_statements, that values representing elapsed time should
be represented as float8 and measured in milliseconds. By using float8,
we are no longer tied to a specific maximum precision of timing data.
(Internally, it's still microseconds, but we could now change that without
needing changes at the SQL level.)
The columns affected are
pg_stat_bgwriter.checkpoint_write_time
pg_stat_bgwriter.checkpoint_sync_time
pg_stat_database.blk_read_time
pg_stat_database.blk_write_time
pg_stat_user_functions.total_time
pg_stat_user_functions.self_time
pg_stat_xact_user_functions.total_time
pg_stat_xact_user_functions.self_time
The first four of these are new in 9.2, so there is no compatibility issue
from changing them. The others require a release note comment that they
are now double precision (and can show a fractional part) rather than
bigint as before; also their underlying statistics functions now match
the column definitions, instead of returning bigint microseconds.
The alternative of disallowing index-only scans in HS operation was
discussed, but the consensus was that it was better to treat marking
a page all-visible as a recovery conflict for snapshots that could still
fail to see XIDs on that page. We may in the future try to soften this,
so that we simply force index scans to do heap fetches in cases where
this may be an issue, rather than throwing a hard conflict.
The previous code could cause a backend crash after BEGIN; SAVEPOINT a;
LOCK TABLE foo (interrupted by ^C or statement timeout); ROLLBACK TO
SAVEPOINT a; LOCK TABLE foo, and might have leaked strong-lock counts
in other situations.
Report by Zoltán Böszörményi; patch review by Jeff Davis.
When using synchronous replication, we waited for the commit record to be
replicated, but if we our transaction didn't write any other WAL records,
that's not required because we don't even flush the WAL locally to disk in
that case. This lead to long waits when committing a transaction that only
modified a temporary table. Bug spotted by Thom Brown.
XLOG_GIN_UPDATE_META_PAGE and XLOG_GIN_DELETE_LISTPAGE records were printed
with a list link field labeled as "blkno", which was confusing, especially
when the link was empty (InvalidBlockNumber). Print the metapage block
number instead, since that's what's actually being updated. We could
include the link values too as a separate field, but not clear it's worth
the trouble.
Back-patch to 8.4 where the dubious code was added.
Initialise ckptXidEpoch from starting checkpoint and maintain the correct
value as we roll forwards. This allows GetNextXidAndEpoch() to return the
correct epoch when executed during recovery. Backpatch to 9.0 when the
problem is first observable by a user.
Bug report from Daniel Farina
It used to be case that lazy vacuum could call this function with only
a shared lock on the buffer, but neither lazy vacuum nor any other
code path does that any more. Simplify the code accordingly and clean
up some related, obsolete comments.
For those variables only used when asserts are enabled, use a new
macro PG_USED_FOR_ASSERTS_ONLY, which expands to
__attribute__((unused)) when asserts are not enabled.
A leaf tuple that we need to delete could get moved as a consequence of an
insertion happening concurrently with the VACUUM scan. If it moves from a
page past the current scan point to a page before, we'll miss it, which is
not acceptable. Hence, when we see a leaf-page REDIRECT that could have
been made since our scan started, chase down the redirection pointer much
as if we were doing a normal index search, and be sure to vacuum every page
it leads to. This fixes the issue because, if the tuple was on page N at
the instant we start our scan, we will surely find it as a consequence of
chasing the redirect from page N, no matter how much it moves around in
between. Problem noted by Takashi Yamamoto.
This patch fixes the other major compatibility-breaking limitation of
SPGiST, that it didn't store anything for null values of the indexed
column, and so could not support whole-index scans or "x IS NULL"
tests. The approach is to create a wholly separate search tree for
the null entries, and use fixed "allTheSame" insertion and search
rules when processing this tree, instead of calling the index opclass
methods. This way the opclass methods do not need to worry about
dealing with nulls.
Catversion bump is for pg_am updates as well as the change in on-disk
format of SPGiST indexes; there are some tweaks in SPGiST WAL records
as well.
Heavily rewritten version of a patch by Oleg Bartunov and Teodor Sigaev.
(The original also stored nulls separately, but it reused GIN code to do
so; which required undesirable compromises in the on-disk format, and
would likely lead to bugs due to the GIN code being required to work in
two very different contexts.)
The original API definition was incapable of supporting whole-index scans
because there was no way to invoke leaf-value reconstruction without
checking any qual conditions. Also, it was inefficient for
multiple-qual-condition scans because value reconstruction got done over
again for each qual condition, and because other internal work in the
consistent functions likewise had to be done for each qual. To fix these
issues, pass the whole scankey array to the opclass consistent functions,
instead of only letting them see one item at a time. (Essentially, the
loop over scankey entries is now inside the consistent functions not
outside them. This makes the consistent functions a bit more complicated,
but not unreasonably so.)
In itself this commit does nothing except save a few cycles in
multiple-qual-condition index scans, since we can't support whole-index
scans on SPGiST indexes until nulls are included in the index. However,
I consider this a must-fix for 9.2 because once we release it will get
very much harder to change the opclass API definition.
This simplifies the code a little bit. The new rule is that to update
XLogCtl->LogwrtResult, you must hold both WALWriteLock and info_lck, whereas
before we had two copies, one that was protected by WALWriteLock and another
protected by info_lck. The code that updates them was already holding both
locks, so merging the two is trivial.
The third copy, XLogCtl->Insert.LogwrtResult, was not totally redundant, it
was used in AdvanceXLInsertBuffer to update the backend-local copy, before
acquiring the info_lck to read the up-to-date value. But the value of that
seems dubious; at best it's saving one spinlock acquisition per completed
WAL page, which is not significant compared to all the other work involved.
And in practice, it's probably not saving even that much.
It's harmless to do full page writes even when not strictly necessary, so
when turning full_page_writes on, we can set the global flag first, and then
call XLogInsert. Likewise, when turning it off, we can write the WAL record
first, and then clear the flag. This way XLogInsert doesn't need any special
handling of the XLOG_FPW_CHANGE record type. XLogInsert is complicated
enough already, so anything we can keep away from there is a good thing.
Actually I don't think the atomicity of the shared memory flag matters,
anyway, because we only write the XLOG_FPW_CHANGE at the end of recovery,
when there are no concurrent WAL insertions going on. But might as well make
it safe, in case we allow changing full_page_writes on the fly in the
future.
Now that we have validate_xlog_location, call it from the previously
existing functions taking xlog locatoins as a string input.
Suggested by Fujii Masao
Comparing two xlog locations are useful for example when calculating
replication lag.
Euler Taveira de Oliveira, reviewed by Fujii Masao, and some cleanups
from me
Previously it was thought that it's impossible as the code stands, because
insertions create buffers as tuples are cascaded downwards, and index
split also creaters buffers eagerly for all halves. But the example from
Jay Levitt demonstrates that it can happen, when the root page is split.
It's in fact OK if the buffer doesn't exist, so we just need to remove the
sanity check. In fact, we've been discussing the possibility of destroying
empty buffers to conserve memory, which would render the sanity check
completely useless anyway.
Fix by Alexander Korotkov
In commit 4016bdef8a I fixed a bunch of
ginxlog.c bugs having to do with not handling XLogReadBuffer failures
correctly. However, in ginRedoUpdateMetapage and ginRedoDeleteListPages,
I unaccountably thought that failure to read the metapage would be
impossible and just put in an elog(PANIC) call. This is of course wrong:
failure is exactly what will happen if the index got dropped (or rebuilt)
between creation of the WAL record and the crash we're trying to recover
from. I believe this explains Nicholas Wilson's recent report of these
errors getting reached.
Also, fix memory leak in forgetIncompleteSplit. This wasn't of much
concern when the code was written, but in a long-running standby server
page split records could be expected to accumulate indefinitely.
Back-patch to 8.4 --- before that, GIN didn't have a metapage.
When "vacuuming" a single btree page by removing LP_DEAD tuples, we are not
actually within a vacuum operation, but rather in an ordinary insertion
process that could well be running concurrently with a vacuum. So clearing
the cycleid is incorrect, and could cause the concurrent vacuum to miss
removing tuples that it needs to remove. This is a longstanding bug
introduced by commit e6284649b9 of
2006-07-25. I believe it explains Maxim Boguk's recent report of index
corruption, and probably some other previously unexplained reports.
In 9.0 and up this is a one-line fix; before that we need to introduce a
flag to tell _bt_delitems what to do.
If tuples were toasted, heap_multi_insert didn't update the ctid on the
original tuples. This caused a failure if there was an after trigger
(including a foreign key), on the table, and a tuple got toasted.
Per off-list report and test case from Ted Phelps
Originally, most of this code assumed that no Postgres backends could be
running concurrently with it, and so no locking could be needed. That
assumption fails in Hot Standby. While it's still true that Hot Standby
backends should never change values like nextXid, they can examine them,
and consistency is important in some cases such as when computing a
snapshot. Therefore, prudence requires that WAL replay code obtain the
relevant locks when modifying such variables, even though it can examine
them without taking a lock. We were following that coding rule in some
places but not all. This commit applies the coding rule uniformly to all
updates of ShmemVariableCache and MultiXactState fields; a search of the
replay routines did not find any other cases that seemed to be at risk.
In addition, this commit fixes a longstanding thinko in replay of NEXTOID
and checkpoint records: we tried to advance nextOid only if it was behind
the value in the WAL record, but the comparison would draw the wrong
conclusion if OID wraparound had occurred since the previous value.
Better to just unconditionally assign the new value, since OID assignment
shouldn't be happening during replay anyway.
The additional locking seems to be more in the nature of future-proofing
than fixing any live bug, so I am not going to back-patch it. The NEXTOID
fix will be back-patched separately.
RestoreBkpBlocks was in the habit of zeroing and refilling the target
buffer; which was perfectly safe when the code was written, but is unsafe
during Hot Standby operation. The reason is that we have coding rules
that allow backends to continue accessing a tuple in a heap relation while
holding only a pin on its buffer. Such a backend could see transiently
zeroed data, if WAL replay had occasion to change other data on the page.
This has been shown to be the cause of bug #6425 from Duncan Rance (who
deserves kudos for developing a sufficiently-reproducible test case) as
well as Bridget Frey's re-report of bug #6200. It most likely explains the
original report as well, though we don't yet have confirmation of that.
To fix, change the code so that only bytes that are supposed to change will
change, even transiently. This actually saves cycles in RestoreBkpBlocks,
since it's not writing the same bytes twice.
Also fix seq_redo, which has the same disease, though it has to work a bit
harder to meet the requirement.
So far as I can tell, no other WAL replay routines have this type of bug.
In particular, the index-related replay routines, which would certainly be
broken if they had to meet the same standard, are not at risk because we
do not have coding rules that allow access to an index page when not
holding a buffer lock on it.
Back-patch to 9.0 where Hot Standby was added.
When testing bits (but not when setting or clearing them), we now
won't check whether the map has been extended. This significantly
improves performance in the case where the visibility map doesn't
exist yet, by avoiding an extra system call per tuple. To make
sure backends notice eventually, send an smgr inval on VM extension.
Dean Rasheed, with minor modifications by me.
When a backend needs to flush the WAL, and someone else is already flushing
the WAL, wait until it releases the WALInsertLock and check if we still need
to do the flush or if the other backend already did the work for us, before
acquiring WALInsertLock. This helps group commit, because when the WAL flush
finishes, all the backends that were waiting for it can be woken up in one
go, and the can all concurrently observe that they're done, rather than
waking them up one by one in a cascading fashion.
This is based on a new LWLock function, LWLockWaitUntilFree(), which has
peculiar semantics. If the lock is immediately free, it grabs the lock and
returns true. If it's not free, it waits until it is released, but then
returns false without grabbing the lock. This is used in XLogFlush(), so
that when the lock is acquired, the backend flushes the WAL, but if it's
not, the backend first checks the current flush location before retrying.
Original patch and benchmarking by Peter Geoghegan and Simon Riggs, although
this patch as committed ended up being very different from that.
Base backup follows recommended procedure, plus goes to great
lengths to ensure that partial page writes are avoided.
Jun Ishizuka and Fujii Masao, with minor modifications
Previously we used ReadRecPtr rather than EndRecPtr, which was
not a serious error but caused pg_stat_replication to report
incorrect replay_location until at least one WAL record is replayed.
Fujii Masao
In commit 7b0d0e9356, I made CLUSTER and
VACUUM FULL try to preserve toast value OIDs from the original toast table
to the new one. However, if we have to copy both live and recently-dead
versions of a row that has a toasted column, those versions may well
reference the same toast value with the same OID. The patch then led to
duplicate-key failures as we tried to insert the toast value twice with the
same OID. (The previous behavior was not very desirable either, since it
would have silently inserted the same value twice with different OIDs.
That wastes space, but what's worse is that the toast values inserted for
already-dead heap rows would not be reclaimed by subsequent ordinary
VACUUMs, since they go into the new toast table marked live not deleted.)
To fix, check if the copied OID already exists in the new toast table, and
if so, assume that it stores the desired value. This is reasonably safe
since the only case where we will copy an OID from a previous toast pointer
is when toast_insert_or_update was given that toast pointer and so we just
pulled the data from the old table; if we got two different values that way
then we have big problems anyway. We do have to assume that no other
backend is inserting items into the new toast table concurrently, but
that's surely safe for CLUSTER and VACUUM FULL.
Per bug #6393 from Maxim Boguk. Back-patch to 9.0, same as the previous
patch.
constructed before acquiring WALInsertLock, which slightly reduces the time
the lock is held. Although I could not measure any benefit in benchmarks,
the code is more readable this way.
Previously, this was hardcoded: we always had 8. Performance testing
shows that isn't enough, especially on big SMP systems, so we allow it
to scale up as high as 32 when there's adequate memory. On the flip
side, when shared_buffers is very small, drop the number of CLOG buffers
down to as little as 4, so that we can start the postmaster even
when very little shared memory is available.
Per extensive discussion with Simon Riggs, Tom Lane, and others on
pgsql-hackers.
In commit e2c2c2e8b1 I made use of nested
list structures to show which clauses went with which index columns, but
on reflection that's a data structure that only an old-line Lisp hacker
could love. Worse, it adds unnecessary complication to the many places
that don't much care which clauses go with which index columns. Revert
to the previous arrangement of flat lists of clauses, and instead add a
parallel integer list of column numbers. The places that care about the
pairing can chase both lists with forboth(), while the places that don't
care just examine one list the same as before.
The only real downside to this is that there are now two more lists that
need to be passed to amcostestimate functions in case they care about
column matching (which btcostestimate does, so not passing the info is not
an option). Rather than deal with 11-argument amcostestimate functions,
pass just the IndexPath and expect the functions to extract fields from it.
That gets us down to 7 arguments which is better than 11, and it seems
more future-proof against likely additions to the information we keep
about an index path.
When a view is marked as a security barrier, it will not be pulled up
into the containing query, and no quals will be pushed down into it,
so that no function or operator chosen by the user can be applied to
rows not exposed by the view. Views not configured with this
option cannot provide robust row-level security, but will perform far
better.
Patch by KaiGai Kohei; original problem report by Heikki Linnakangas
(in October 2009!). Review (in earlier versions) by Noah Misch and
others. Design advice by Tom Lane and myself. Further review and
cleanup by me.
smgrdounlink takes care to not throw an ERROR if it fails to unlink
something, but that caution was rendered useless by commit
3396000684, which put an smgrexists call in
front of it; smgrexists *does* throw error if anything looks funny, such
as getting a permissions error from trying to open the file. If that
happens post-commit, you get a PANIC, and what's worse the same logic
appears in the WAL replay code, so the database even fails to restart.
Restore the intended behavior by removing the smgrexists call --- it isn't
accomplishing anything that we can't do better by adjusting mdunlink's
ideas of whether it ought to warn about ENOENT or not.
Per report from Joseph Shraibman of unrecoverable crash after trying to
drop a table whose FSM fork had somehow gotten chmod'd to 000 permissions.
Backpatch to 8.4, where the bogus coding was introduced.
This adds support for the more or less SQL-conforming USAGE privilege
on types and domains. The intent is to be able restrict which users
can create dependencies on types, which restricts the way in which
owners can alter types.
reviewed by Yeb Havinga
On reflection, the original name seems way too generic for a global
symbol. A quick check shows this is the only exported function name
in SP-GiST that doesn't begin with "spg" or contain "SpGist", so the
rest of them seem all right.
Operator classes can specify whether or not they support this; this
preserves the flexibility to use lossy representations within an index.
In passing, move constant data about a given index into the rd_amcache
cache area, instead of doing fresh lookups each time we start an index
operation. This is mainly to try to make sure that spgcanreturn() has
insignificant cost; I still don't have any proof that it matters for
actual index accesses. Also, get rid of useless copying of FmgrInfo
pointers; we can perfectly well use the relcache's versions in-place.
The need for this was debated when we put in the index-only-scan feature,
but at the time we had no near-term expectation of having AMs that could
support such scans for only some indexes; so we kept it simple. However,
the SP-GiST AM forces the issue, so let's fix it.
This patch only installs the new API; no behavior actually changes.
SP-GiST is comparable to GiST in flexibility, but supports non-balanced
partitioned search structures rather than balanced trees. As described at
PGCon 2011, this new indexing structure can beat GiST in both index build
time and query speed for search problems that it is well matched to.
There are a number of areas that could still use improvement, but at this
point the code seems committable.
Teodor Sigaev and Oleg Bartunov, with considerable revisions by Tom Lane
Removing this bit from xl_info allows us to restore the old limit of four
(not three) separate pages touched by a WAL record, which is needed for the
upcoming SP-GiST feature, and will likely be useful elsewhere in future.
When we implemented XLR_BKP_REMOVABLE in 2007, we had to do it like that
because no special WAL-visible action was taken when starting a backup.
However, now we force a segment switch when starting a backup, so a
compressing WAL archiver (such as pglesslog) that uses the state shown in
the current page header will not be fooled as to removability of backup
blocks. The only downside is that the archiver will not return to
compressing mode for up to one WAL page after the backup is over, which is
a small price to pay for getting back the extra xl_info bit. In any case
the archiver could look for XLOG_BACKUP_END records if it thought it was
worth the trouble to do so.
Bump XLOG_PAGE_MAGIC since this is effectively a change in WAL format.
we don't reach consistency before replaying all of the WAL. Rename the
variable to reachedConsistency, to make its intention clearer.
In master, that was an active bug because of the recent patch to
immediately PANIC if a reference to a missing page is found in WAL after
reaching consistency, as Tom Lane's test case demonstrated. In 9.1 and 9.0,
the only consequence was a misleading "consistent recovery state reached at
%X/%X" message in the log at the beginning of crash recovery (the database
is not consistent at that point yet). In 8.4, the log message was not
printed in crash recovery, even though there was a similar
reachedMinRecoveryPoint local variable that was also set early. So,
backpatch to 9.1 and 9.0.
This patch creates an API whereby a btree index opclass can optionally
provide non-SQL-callable support functions for sorting. In the initial
patch, we only use this to provide a directly-callable comparator function,
which can be invoked with a bit less overhead than the traditional
SQL-callable comparator. While that should be of value in itself, the real
reason for doing this is to provide a datatype-extensible framework for
more aggressive optimizations, as in Peter Geoghegan's recent work.
Robert Haas and Tom Lane
invalid-page hash table, PANIC immediately. Immediate PANIC is much better
than waiting for end-of-recovery, which is what we did before, because the
end-of-recovery might not come until months later if this is a standby
server.
Also refrain from creating a restartpoint if there are invalid-page entries
in the hash table. Restarting recovery from such a restartpoint would not
see the invalid references, and wouldn't be able to cross-check them when
consistency is reached. That wouldn't matter when things are going smoothly,
but the more sanity checks you have the better.
Fujii Masao
In the previous coding, callers were faced with an awkward choice:
look up the name, do permissions checks, and then lock the table; or
look up the name, lock the table, and then do permissions checks.
The first choice was wrong because the results of the name lookup
and permissions checks might be out-of-date by the time the table
lock was acquired, while the second allowed a user with no privileges
to interfere with access to a table by users who do have privileges
(e.g. if a malicious backend queues up for an AccessExclusiveLock on
a table on which AccessShareLock is already held, further attempts
to access the table will be blocked until the AccessExclusiveLock
is obtained and the malicious backend's transaction rolls back).
To fix, allow callers of RangeVarGetRelid() to pass a callback which
gets executed after performing the name lookup but before acquiring
the relation lock. If the name lookup is retried (because
invalidation messages are received), the callback will be re-executed
as well, so we get the best of both worlds. RangeVarGetRelid() is
renamed to RangeVarGetRelidExtended(); callers not wishing to supply
a callback can continue to invoke it as RangeVarGetRelid(), which is
now a macro. Since the only one caller that uses nowait = true now
passes a callback anyway, the RangeVarGetRelid() macro defaults nowait
as well. The callback can also be used for supplemental locking - for
example, REINDEX INDEX needs to acquire the table lock before the index
lock to reduce deadlock possibilities.
There's a lot more work to be done here to fix all the cases where this
can be a problem, but this commit provides the general infrastructure
and fixes the following specific cases: REINDEX INDEX, REINDEX TABLE,
LOCK TABLE, and and DROP TABLE/INDEX/SEQUENCE/VIEW/FOREIGN TABLE.
Per discussion with Noah Misch and Alvaro Herrera.
Use of a randomly chosen large value was never exactly graceful, and
now that there are penalty functions that are intentionally using infinity,
it doesn't seem like a good idea for null-vs-not-null to be using something
less.
A simple thinko in ginRedoUpdateMetapage, namely failing to increment a
loop counter, led to inserting records into the last pending-list page in
the wrong order (the opposite of that intended). So far as I can tell,
this would not upset the code that eventually flushes pending items into
the main part of the GIN index. But it did break the code that searched
the pending list for matches, resulting in transient failure to find
matching entries during index lookups, as illustrated in bug #6307 from
Maksym Boguk.
Back-patch to 8.4 where the incorrect code was introduced.
This speeds up snapshot-taking and reduces ProcArrayLock contention.
Also, the PGPROC (and PGXACT) structures used by two-phase commit are
now allocated as part of the main array, rather than in a separate
array, and we keep ProcArray sorted in pointer order. These changes
are intended to minimize the number of cache lines that must be pulled
in to take a snapshot, and testing shows a substantial increase in
performance on both read and write workloads at high concurrencies.
Pavan Deolasee, Heikki Linnakangas, Robert Haas
When wal_level = 'hot_standby' we touched the last page of the
relation during a VACUUM, even if nothing else had happened.
That would alter the LSN of the last block and set the mtime
of the relation file unnecessarily. Noted by Thom Brown.
Previously we waited for wal_writer_delay before flushing WAL. Now
we also wake WALWriter as soon as a WAL buffer page has filled.
Significant effect observed on performance of asynchronous commits
by Robert Haas, attributed to the ability to set hint bits on tuples
earlier and so reducing contention caused by clog lookups.
Forgot to call RestoreBkpBlocks() in the redo-function, as pointed out by
Simon Riggs. In redo of a regular heap insert, it's taken care of in
heap_redo(), but this new record type uses the heap2 RM, and heap2_redo()
does not take care of that for you.
Also, failed to reset the vmbuffer and all_visibile_cleared local variables
after switching to a new buffer.
This greatly reduces the WAL volume, especially when the table is narrow.
The overhead of locking the heap page is also reduced. Reduced WAL traffic
also makes it scale a lot better, if you run multiple COPY processes at
the same time.
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
This assumption can be wrong when the toaster is passed a raw on-disk
tuple, because the tuple might pre-date an ALTER TABLE ADD COLUMN operation
that added columns without rewriting the table. In such a case the tuple's
natts value is smaller than what we expect from the tuple descriptor, and
so its t_hoff value could be smaller too. In fact, the tuple might not
have a null bitmap at all, and yet our current opinion of it is that it
contains some trailing nulls.
In such a situation, toast_insert_or_update did the wrong thing, because
to save a few lines of code it would use the old t_hoff value as the offset
where heap_fill_tuple should start filling data. This did not leave enough
room for the new nulls bitmap, with the result that the first few bytes of
data could be overwritten with null flag bits, as in a recent report from
Hubert Depesz Lubaczewski.
The particular case reported requires ALTER TABLE ADD COLUMN followed by
CREATE TABLE AS SELECT * FROM ... or INSERT ... SELECT * FROM ..., and
further requires that there be some out-of-line toasted fields in one of
the tuples to be copied; else we'll not reach the troublesome code.
The problem can only manifest in this form in 8.4 and later, because
before commit a77eaa6a95, CREATE TABLE AS or
INSERT/SELECT wouldn't result in raw disk tuples getting passed directly
to heap_insert --- there would always have been at least a junkfilter in
between, and that would reconstitute the tuple header with an up-to-date
t_natts and hence t_hoff. But I'm backpatching the tuptoaster change all
the way anyway, because I'm not convinced there are no older code paths
that present a similar risk.
If we have an inequality key that constrains the other end of the index,
it doesn't directly help us in doing the initial positioning ... but it
does imply a NOT NULL constraint on the index column. If the index stores
nulls at this end, we can use the implied NOT NULL condition for initial
positioning, just as if it had been stated explicitly. This avoids wasting
time when there are a lot of nulls in the column. This is the reverse of
the examples given in bugs #6278 and #6283, which were about failing to
stop early when we encounter nulls at the end of the indexscan.
As pointed out by Naoya Anzai, my previous try at this was a few bricks
shy of a load, because I had forgotten that the initial-positioning logic
might not try to skip over nulls at the end of the index the scan will
start from. We ought to fix that, because it represents an unnecessary
inefficiency, but first let's get the scan-stop logic back to a safe
state. With this patch, we preserve the performance benefit requested
in bug #6278 for the case of scanning forward into NULLs (in a NULLS
LAST index), but the reverse case of scanning backward across NULLs
when there's no suitable initial-positioning qual is still inefficient.
Previously, we skipped a checkpoint if no WAL had been written since
last checkpoint, though this does not appear in user documentation.
As of now, we skip a checkpoint until we have written at least one
enough WAL to switch the next WAL file. This greatly reduces the
level of activity and number of WAL messages generated by a very
low activity server. This is safe because the purpose of a checkpoint
is to act as a starting place for a recovery, in case of crash.
This patch maintains minimal WAL volume for replay in case of crash,
thus maintaining very low crash recovery time.
There was a timing window between when oldestActiveXid was derived
and when it should have been derived that only shows itself under
heavy load. Move code around to ensure correct timing of derivation.
No change to StartupSUBTRANS() code, which is where this failed.
Bug report by Chris Redekop
If a tuple in a syscache contains an out-of-line toasted field, and we
try to fetch that field shortly after some other transaction has committed
an update or deletion of the tuple, there is a race condition: vacuum
could come along and remove the toast tuples before we can fetch them.
This leads to transient failures like "missing chunk number 0 for toast
value NNNNN in pg_toast_2619", as seen in recent reports from Andrew
Hammond and Tim Uckun.
The design idea of syscache is that access to stale syscache entries
should be prevented by relation-level locks, but that fails for at least
two cases where toasted fields are possible: ANALYZE updates pg_statistic
rows without locking out sessions that might want to plan queries on the
same table, and CREATE OR REPLACE FUNCTION updates pg_proc rows without
any meaningful lock at all.
The least risky fix seems to be an idea that Heikki suggested when we
were dealing with a related problem back in August: forcibly detoast any
out-of-line fields before putting a tuple into syscache in the first place.
This avoids the problem because at the time we fetch the parent tuple from
the catalog, we should be holding an MVCC snapshot that will prevent
removal of the toast tuples, even if the parent tuple is outdated
immediately after we fetch it. (Note: I'm not convinced that this
statement holds true at every instant where we could be fetching a syscache
entry at all, but it does appear to hold true at the times where we could
fetch an entry that could have a toasted field. We will need to be a bit
wary of adding toast tables to low-level catalogs that don't have them
already.) An additional benefit is that subsequent uses of the syscache
entry should be faster, since they won't have to detoast the field.
Back-patch to all supported versions. The problem is significantly harder
to reproduce in pre-9.0 releases, because of their willingness to flush
every entry in a syscache whenever the underlying catalog is vacuumed
(cf CatalogCacheFlushRelation); but there is still a window for trouble.
The existing scan-direction-sensitive tests were overly complex, and
failed to stop the scan in cases where it's perfectly legitimate to do so.
Per bug #6278 from Maksym Boguk.
Back-patch to 8.3, which is as far back as the patch applies easily.
Doesn't seem worth sweating over a relatively minor performance issue in
8.2 at this late date. (But note that this was a performance regression
from 8.1 and before, so 8.2 is being left as an outlier.)
A transaction can export a snapshot with pg_export_snapshot(), and then
others can import it with SET TRANSACTION SNAPSHOT. The data does not
leave the server so there are not security issues. A snapshot can only
be imported while the exporting transaction is still running, and there
are some other restrictions.
I'm not totally convinced that we've covered all the bases for SSI (true
serializable) mode, but it works fine for lesser isolation modes.
Joachim Wieland, reviewed by Marko Tiikkaja, and rather heavily modified
by Tom Lane
This is merely an exercise in satisfying pedants, not a bug fix, because
in every case we were checking for failure later with ferror(), or else
there was nothing useful to be done about a failure anyway. Document
the latter cases.
In general the data returned by an index-only scan should have the
datatypes originally computed by FormIndexDatum. If the index opclasses
use "storage" datatypes different from their input datatypes, the scan
tuple will not have the same rowtype attributed to the index; but we had
a hard-wired assumption that that was true in nodeIndexonlyscan.c. We'd
already hacked around the issue for the one case where the types are
different in btree indexes (btree name_ops), but this would definitely
come back to bite us if we ever implement index-only scans in GiST.
To fix, require the index AM to explicitly provide the tupdesc for the
tuple it is returning. btree can just pass back the index's tupdesc, but
GiST will have to work harder when and if it supports index-only scans.
I had previously proposed fixing this by allowing the index AM to fill the
scan tuple slot directly; but on reflection that seemed like a module
layering violation, since TupleTableSlots are creatures of the executor.
At least in the btree case, it would also be less efficient, since the
tuple deconstruction work would occur even for rows later found to be
invisible to the scan's snapshot.
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.
It's been bothering me for several days that pretending that the cstring
data stored in a btree name_ops column is really a "name" Datum could lead
to reading past the end of memory. However, given the current memory
layout used for index-only scans in the btree code, a crash is in fact not
possible. Document that so we don't break it. I have not thought of any
other solutions that aren't fairly ugly too, and most of them lose the
functionality of index-only scans on name columns altogether, so this seems
like the way to go.
The original idea of this patch was to make box picksplit run faster, by
eliminating unnecessary palloc() overhead, but that was obsoleted by the new
double-sorting split algorithm that doesn't call these functions so heavily
anymore. Nevertheless, the code looks better this way.
Original patch by me, reviewed and tidied up after the double-sorting patch
by Kevin Grittner.
We copy all the matched tuples off the page during _bt_readpage, instead of
expensively re-locking the page during each subsequent tuple fetch. This
costs a bit more local storage, but not more than 2*BLCKSZ worth, and the
reduction in LWLock traffic is certainly worth that. What's more, this
lets us get rid of the API wart in the original patch that said an index AM
could randomly decline to supply an index tuple despite having asserted
pg_am.amcanreturn. That will be important for future improvements in the
index-only-scan feature, since the executor will now be able to rely on
having the index data available.
When a btree index contains all columns required by the query, and the
visibility map shows that all tuples on a target heap page are
visible-to-all, we don't need to fetch that heap page. This patch depends
on the previous patches that made the visibility map reliable.
There's a fair amount left to do here, notably trying to figure out a less
chintzy way of estimating the cost of an index-only scan, but the core
functionality seems ready to commit.
Robert Haas and Ibrar Ahmed, with some previous work by Heikki Linnakangas.
Previously, the code assumed that the only possible action to take was
to delete files behind a certain cutoff point. The async notify code
was already a crock: it used a different "pagePrecedes" function for
truncation than for regular operation. By allowing it to pass a
callback to SlruScanDirectory it can do cleanly exactly what it needs to
do.
The clog.c code also had its own use for SlruScanDirectory, which is
made a bit simpler with this.
This patch has two distinct purposes: to report multiple problems in
postgresql.conf rather than always bailing out after the first one,
and to change the policy for whether changes are applied when there are
unrelated errors in postgresql.conf.
Formerly the policy was to apply no changes if any errors could be
detected, but that had a significant consistency problem, because in some
cases specific values might be seen as valid by some processes but invalid
by others. This meant that the latter processes would fail to adopt
changes in other parameters even though the former processes had done so.
The new policy is that during SIGHUP, the file is rejected as a whole
if there are any errors in the "name = value" syntax, or if any lines
attempt to set nonexistent built-in parameters, or if any lines attempt
to set custom parameters whose prefix is not listed in (the new value of)
custom_variable_classes. These tests should always give the same results
in all processes, and provide what seems a reasonably robust defense
against loading values from badly corrupted config files. If these tests
pass, all processes will apply all settings that they individually see as
good, ignoring (but logging) any they don't.
In addition, the postmaster does not abandon reading a configuration file
after the first syntax error, but continues to read the file and report
syntax errors (up to a maximum of 100 syntax errors per file).
The postmaster will still refuse to start up if the configuration file
contains any errors at startup time, but these changes allow multiple
errors to be detected and reported before quitting.
Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?)
with some additional hacking by Tom Lane
pg_trgm was already doing this unofficially, but the implementation hadn't
been thought through very well and leaked memory. Restructure the core
GiST code so that it actually works, and document it. Ordinarily this
would have required an extra memory context creation/destruction for each
GiST index search, but I was able to avoid that in the normal case of a
non-rescanned search by finessing the handling of the RBTree. It used to
have its own context always, but now shares a context with the
scan-lifespan data structures, unless there is more than one rescan call.
This should make the added overhead unnoticeable in typical cases.
In hio.c, document how we avoid deadlock with respect to visibility map
buffer locks. In visibilitymap.c, update the LOCKING section of the
file header comment.
Both oversights noted by Heikki Linnakangas.
In REPEATABLE READ (nee SERIALIZABLE) mode, an attempt to do
GetTransactionSnapshot() between AbortTransaction and CleanupTransaction
failed, because GetTransactionSnapshot would recompute the transaction
snapshot (which is already wrong, given the isolation mode) and then
re-register it in the TopTransactionResourceOwner, leading to an Assert
because the TopTransactionResourceOwner should be empty of resources after
AbortTransaction. This is the root cause of bug #6218 from Yamamoto
Takashi. While changing plancache.c to avoid requesting a snapshot when
handling a ROLLBACK masks the problem, I think this is really a snapmgr.c
bug: it's lower-level than the resource manager mechanism and should not be
shutting itself down before we unwind resource manager resources. However,
just postponing the release of the transaction snapshot until cleanup time
didn't work because of the circular dependency with
TopTransactionResourceOwner. Fix by managing the internal reference to
that snapshot manually instead of depending on TopTransactionResourceOwner.
This saves a few cycles as well as making the module layering more
straightforward. predicate.c's dependencies on TopTransactionResourceOwner
go away too.
I think this is a longstanding bug, but there's no evidence that it's more
than a latent bug, so it doesn't seem worth any risk of back-patching.
As observed by Heikki, we need not conflict on heap page locks during an
insert; heap page locks are only aggregated tuple locks, they don't imply
locking "gaps" as index page locks do. So we can avoid some unnecessary
conflicts, and also do the SSI check while not holding exclusive lock on
the target buffer.
Kevin Grittner, reviewed by Jeff Davis. Back-patch to 9.1.
This oversight led to a massive memory leak --- upwards of 10KB per tuple
--- during creation-time verification of an exclusion constraint based on a
GIST index. In most other scenarios it'd just be a leak of 10KB that would
be recovered at end of query, so not too significant; though perhaps the
leak would be noticeable in a situation where a GIST index was being used
in a nestloop inner indexscan. In any case, it's a real leak of long
standing, so patch all supported branches. Per report from Harald Fuchs.
queuedForEmptying flag correctly on buffer when adding it to the queue.
Also, don't add buffer to the queue if it's there already. These were
harmless oversights; failing to set the flag just means that a buffer might
get added to the queue twice if more tuples are added to it (although that
can't actually happen at this point because all the upper buffers have
already been emptied), and having the same buffer twice in the emptying
queue is harmless. But better be tidy.
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.
When building a GiST index that doesn't fit in cache, buffers are attached
to some internal nodes in the index. This speeds up the build by avoiding
random I/O that would otherwise be needed to traverse all the way down the
tree to the find right leaf page for tuple.
Alexander Korotkov
walsender.h should depend on xlog.h, not vice versa. (Actually, the
inclusion was circular until a couple hours ago, which was even sillier;
but Bruce broke it in the expedient rather than logically correct
direction.) Because of that poor decision, plus blind application of
pgrminclude, we had a situation where half the system was depending on
xlog.h to include such unrelated stuff as array.h and guc.h. Clean up
the header inclusion, and manually revert a lot of what pgrminclude had
done so things build again.
This episode reinforces my feeling that pgrminclude should not be run
without adult supervision. Inclusion changes in header files in particular
need to be reviewed with great care. More generally, it'd be good if we
had a clearer notion of module layering to dictate which headers can sanely
include which others ... but that's a big task for another day.
This requires adjusting the API for syscache callback functions: they now
get a hash value, not a TID, to identify the target tuple. Most of them
weren't paying any attention to that argument anyway, but plancache did
require a small amount of fixing.
Also, improve performance a trifle by avoiding sending duplicate inval
messages when a heap_update isn't changing the catcache lookup columns.
Bruce Momjian
Tom Lane
Simon Riggs
Robert Haas
Heikki Linnakangas
Peter Eisentraut
Magnus Hagander
Alvaro Herrera