When pg_dump retrieves the list of database objects and performs the
data dump, there was possibility that objects are replaced with others
of the same name, such as views, and access them. This vulnerability
could result in code execution with superuser privileges during the
pg_dump process.
This issue can arise when dumping data of sequences, foreign
tables (only 13 or later), or tables registered with a WHERE clause in
the extension configuration table.
To address this, pg_dump now utilizes the newly introduced
restrict_nonsystem_relation_kind GUC parameter to restrict the
accesses to non-system views and foreign tables during the dump
process. This new GUC parameter is added to back branches too, but
these changes do not require cluster recreation.
Back-patch to all supported branches.
Reviewed-by: Noah Misch
Security: CVE-2024-7348
Backpatch-through: 12
This reverts commit 4853630537.
Some buildfarm animals are failing with "cannot change
"client_encoding" during a parallel operation". It looks like
assign_client_encoding is unhappy at being asked to roll back a
client_encoding setting after a parallel worker encounters a
failure. There must be more to it though: why didn't I see this
during local testing? In any case, it's clear that moving the
RestoreGUCState() call is not as side-effect-free as I thought.
Given that the bug f5f30c22e intended to fix has gone unreported
for years, it's not something that's urgent to fix; I'm not
willing to risk messing with it further with only days to our
next release wrap.
Parallel workers failed after a sequence like
BEGIN;
CREATE USER foo;
SET SESSION AUTHORIZATION foo;
because check_session_authorization could not see the uncommitted
pg_authid row for "foo". This is because we ran RestoreGUCState()
in a separate transaction using an ordinary just-created snapshot.
The same disease afflicts any other GUC that requires catalog lookups
and isn't forgiving about the lookups failing.
To fix, postpone RestoreGUCState() into the worker's main transaction
after we've set up a snapshot duplicating the leader's. This affects
check_transaction_isolation and check_transaction_deferrable, which
think they should only run during transaction start. Make them
act like check_transaction_read_only, which already knows it should
silently accept the value when InitializingParallelWorker.
Per bug #18545 from Andrey Rachitskiy. Back-patch to all
supported branches, because this has been wrong for awhile.
Discussion: https://postgr.es/m/18545-feba138862f19aaa@postgresql.org
pg_size_pretty(bigint) would return the value in bytes rather than PB
for the smallest-most bigint value. This happened due to an incorrect
assumption that the absolute value of -9223372036854775808 could be
stored inside a signed 64-bit type.
Here we fix that by instead storing that value in an unsigned 64-bit type.
This bug does exist in versions prior to 15 but the code there is
sufficiently different and the bug seems sufficiently non-critical that
it does not seem worth risking backpatching further.
Author: Joseph Koshakow <koshy44@gmail.com>
Discussion: https://postgr.es/m/CAAvxfHdTsMZPWEHUrZ=h3cky9Ccc3Mtx2whUHygY+ABP-mCmUw@mail.gmail.com
Backpatch-through: 15
Commit 453c468737 introduced a use of strerror() into libpq, but that
is not thread-safe. Fix by using strerror_r() instead.
In passing, update some of the code comments added by 453c468737, as
we have learned more about the reason for the change in OpenSSL that
started this.
Reviewed-by: Daniel Gustafsson <daniel@yesql.se>
Discussion: Discussion: https://postgr.es/m/b6fb018b-f05c-4afd-abd3-318c649faf18@highgo.ca
OpenSSL supports two types of session tickets for TLSv1.3, stateless
and stateful. The option we've used only turns off stateless tickets
leaving stateful tickets active. Use the new API introduced in 1.1.1
to disable all types of tickets.
Backpatch to all supported versions.
Reviewed-by: Heikki Linnakangas <hlinnaka@iki.fi>
Reported-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/20240617173803.6alnafnxpiqvlh3g@awork3.anarazel.de
Backpatch-through: v12
We don't allow inheritance parents as partitions, and have checks to
prevent this; but if a table _was_ in the past an inheritance parents
and all their children are removed, the pg_class.relhassubclass flag
may remain set, which confuses the partition pruning code (most
obviously, it results in an assertion failure; in production builds it
may be worse.)
Fix by resetting relhassubclass on attach.
Backpatch to all supported versions.
Reported-by: Alexander Lakhin <exclusion@gmail.com>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://postgr.es/m/18550-d5e047e9a897a889@postgresql.org
When provided an empty initial array, array_set_slice() fails to
check for overflow when computing the new array's dimensions.
While such overflows are ordinarily caught by ArrayGetNItems(),
commands with the following form are accepted:
INSERT INTO t (i[-2147483648:2147483647]) VALUES ('{}');
To fix, perform the hazardous computations using overflow-detecting
arithmetic routines. As with commit 18b585155a, the added test
cases generate errors that include a platform-dependent value, so
we again use psql's VERBOSITY parameter to suppress printing the
message text.
Reported-by: Alexander Lakhin
Author: Joseph Koshakow
Reviewed-by: Jian He
Discussion: https://postgr.es/m/31ad2cd1-db94-bdb3-f91a-65ffdb4bef95%40gmail.com
Backpatch-through: 12
If a view has some updatable and some non-updatable columns, we failed
to verify updatability of any columns for which an INSERT or UPDATE
on the view explicitly specifies a DEFAULT item (unless the view has
a declared default for that column, which is rare anyway, and one
would almost certainly not write one for a non-updatable column).
This would lead to an unexpected "attribute number N not found in
view targetlist" error rather than the intended error.
Per bug #18546 from Alexander Lakhin. This bug is old, so back-patch
to all supported branches.
Discussion: https://postgr.es/m/18546-84a292e759a9361d@postgresql.org
If vacuum fails to remove a tuple with xmax older than
VacuumCutoffs->OldestXmin and younger than GlobalVisState->maybe_needed,
it will loop infinitely in lazy_scan_prune(), which compares tuples'
visibility information to OldestXmin.
Starting in version 14, which uses GlobalVisState for visibility testing
during pruning, it is possible for GlobalVisState->maybe_needed to
precede OldestXmin if maybe_needed is forced to go backward while vacuum
is running. This can happen if a disconnected standby with a running
transaction older than VacuumCutoffs->OldestXmin reconnects to the
primary after vacuum initially calculates GlobalVisState and OldestXmin.
Fix this by having vacuum always remove tuples older than OldestXmin
during pruning. This is okay because the standby won't replay the tuple
removal until the tuple is removable. Thus, the worst that can happen is
a recovery conflict.
Fixes BUG# 17257
Back-patched in versions 14-17
Author: Melanie Plageman
Reviewed-by: Noah Misch, Peter Geoghegan, Robert Haas, Andres Freund, and Heikki Linnakangas
Discussion: https://postgr.es/m/CAAKRu_Y_NJzF4-8gzTTeaOuUL3CcGoXPjXcAHbTTygT8AyVqag%40mail.gmail.com
Before this change guc_var_compare() cast the input arguments to
const struct config_generic *. That's not quite right however, as the input
on one side is often just a char * on one side.
Instead just use char *, the first field in config_generic.
This fixes a -Warray-bounds warning with some versions of gcc. While the
warning is only known to be triggered for <= 15, the issue the warning points
out seems real, so apply the fix everywhere.
Author: Nazir Bilal Yavuz <byavuz81@gmail.com>
Reported-by: Erik Rijkers <er@xs4all.nl>
Suggested-by: Andres Freund <andres@anarazel.de>
Discussion: https://postgr.es/m/a74a1a0d-0fd2-3649-5224-4f754e8f91aa%40xs4all.nl
checkWellFormedRecursion would issue "missing recursive reference"
if a WITH RECURSIVE query contained a single self-reference but
that self-reference was inside a top-level WITH, ORDER BY, LIMIT,
etc, rather than inside the second arm of the UNION as expected.
We already intended to throw more-on-point errors for such cases,
but those error checks must be done before examining the UNION arm
in order to have the desired results. So this patch need only
move some code (and improve the comments).
Per bug #18536 from Alexander Lakhin. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18536-0a342ec07901203e@postgresql.org
ANALYZE sets relhassubclass=f when a partitioned table no longer has
partitions. An ANALYZE doing that proceeded to apply the inplace update
of pg_class.reltuples to the old pg_class tuple instead of the new
tuple, losing that reltuples=0 change if the ANALYZE committed.
Non-partitioning inheritance trees were unaffected. Back-patch to v14,
where commit 375aed36ad introduced
maintenance of partitioned table pg_class.reltuples.
Reported by Alexander Lakhin.
Discussion: https://postgr.es/m/a295b499-dcab-6a99-c06e-01cf60593344@gmail.com
Winsock only signals an FD_CLOSE event once if the other end of the
socket shuts down gracefully. Because each WaitLatchOrSocket() call
constructs and destroys a new event handle every time, with unlucky
timing we can lose it and hang. We get away with this only if the other
end disconnects non-gracefully, because FD_CLOSE is repeatedly signaled
in that case.
To fix this design flaw in our Windows socket support fundamentally,
we'd probably need to rearchitect it so that a single event handle
exists for the lifetime of a socket, or switch to completely different
multiplexing or async I/O APIs. That's going to be a bigger job
and probably wouldn't be back-patchable.
This brute force kludge closes the race by explicitly polling with
MSG_PEEK before sleeping.
Back-patch to all supported releases. This should hopefully clear up
some random build farm and CI hang failures reported over the years. It
might also allow us to try using graceful shutdown in more places again
(reverted in commit 29992a6) to fix instability in the transmission of
FATAL error messages, but that isn't done by this commit.
Reported-by: Tom Lane <tgl@sss.pgh.pa.us>
Tested-by: Alexander Lakhin <exclusion@gmail.com>
Discussion: https://postgr.es/m/176008.1715492071%40sss.pgh.pa.us
When a partitioned table has an index that doesn't support a constraint,
but a partition has an equivalent index that does, then a DETACH
operation would misbehave: a crash in assertion-enabled systems (because
we fail to find the constraint in the parent that we expect to), or a
broken coninhcount value (-1) in production systems (because we blindly
believe that we've successfully detached the parent).
While we should reject an ATTACH of a partition with such an index, we
have failed to do so in existing releases, so adding an error in stable
releases might break the (unlikely) existing applications that rely on
this behavior. At this point I don't even want to reject them in
master, because it'd break pg_upgrade if such databases exist, and there
would be no easy way to fix existing databases without expensive index
rebuilds.
(Later on we could add ALTER TABLE ... ADD CONSTRAINT USING INDEX to
partitioned tables, which would allow the user to fix such patterns. At
that point we could add more restrictions to prevent the problem from
its root.)
Also, add a test case that leaves one table in this condition, so that
we can verify that pg_upgrade continues to work if we later decide to
change the policy on the master branch.
Backpatch to all supported branches.
Co-authored-by: Tender Wang <tndrwang@gmail.com>
Reported-by: Alexander Lakhin <exclusion@gmail.com>
Reviewed-by: Tender Wang <tndrwang@gmail.com>
Reviewed-by: Michael Paquier <michael@paquier.xyz>
Discussion: https://postgr.es/m/18500-62948b6fe5522f56@postgresql.org
When creating and initializing a logical slot, the restart_lsn is set
to the latest WAL insertion point (or the latest replay point on
standbys). Subsequently, WAL records are decoded from that point to
find the start point for extracting changes in the
DecodingContextFindStartpoint() function. Since the initial
restart_lsn could be in the middle of a transaction, the start point
must be a consistent point where we won't see the data for partial
transactions.
Previously, when not building a full snapshot, serialized snapshots
were restored, and the SnapBuild jumps to the consistent state even
while finding the start point. Consequently, the slot's restart_lsn
and confirmed_flush could be set to the middle of a transaction. This
could lead to various unexpected consequences. Specifically, there
were reports of logical decoding decoding partial transactions, and
assertion failures occurred because only subtransactions were decoded
without decoding their top-level transaction until decoding the commit
record.
To resolve this issue, the changes prevent restoring the serialized
snapshot and jumping to the consistent state while finding the start
point.
On v17 and HEAD, a flag indicating whether snapshot restores should be
skipped has been added to the SnapBuild struct, and SNAPBUILD_VERSION
has been bumpded.
On backbranches, the flag is stored in the LogicalDecodingContext
instead, preserving on-disk compatibility.
Backpatch to all supported versions.
Reported-by: Drew Callahan
Reviewed-by: Amit Kapila, Hayato Kuroda
Discussion: https://postgr.es/m/2444AA15-D21B-4CCE-8052-52C7C2DAFE5C%40amazon.com
Backpatch-through: 12
This back-patches HEAD commits 066e8ac6e, 6082b3d5d, e7192486d,
and 896cd266f into supported branches. Changes:
* Use xmlAddChildList not xmlAddChild in XMLSERIALIZE
(affects v16 and up only). This was a flat-out coding mistake
that we got away with due to lax checking in previous versions
of xmlAddChild.
* Use xmlParseInNodeContext not xmlParseBalancedChunkMemory.
This is to dodge a bug in xmlParseBalancedChunkMemory in libxm2
releases 2.13.0-2.13.2. While that bug is now fixed upstream and
will probably never be seen in any production-oriented distro, it is
currently a problem on some more-bleeding-edge-friendly platforms.
* Suppress "chunk is not well balanced" errors from libxml2,
unless it is the only error. This eliminates an error-reporting
discrepancy between 2.13 and older releases. This error is
almost always redundant with previous errors, if not flat-out
inappropriate, which is why 2.13 changed the behavior and why
nobody's likely to miss it.
Erik Wienhold and Tom Lane, per report from Frank Streitzig.
Discussion: https://postgr.es/m/trinity-b0161630-d230-4598-9ebc-7a23acdb37cb-1720186432160@3c-app-gmx-bap25
Discussion: https://postgr.es/m/trinity-361ba18b-541a-4fe7-bc63-655ae3a7d599-1720259822452@3c-app-gmx-bs01
The numeric round() and trunc() functions clamp the scale argument to
the range between +/- NUMERIC_MAX_RESULT_SCALE (2000), which is much
smaller than the actual allowed range of type numeric. As a result,
they return incorrect results when asked to round/truncate more than
2000 digits before or after the decimal point.
Fix by using the correct upper and lower scale limits based on the
actual allowed (and documented) range of type numeric.
While at it, use the new NUMERIC_WEIGHT_MAX constant instead of
SHRT_MAX in all other overflow checks, and fix a comment thinko in
power_var() introduced by e54a758d24 -- the minimum value of
ln_dweight is -NUMERIC_DSCALE_MAX (-16383), not -SHRT_MAX, though this
doesn't affect the point being made in the comment, that the resulting
local_rscale value may exceed NUMERIC_MAX_DISPLAY_SCALE (1000).
Back-patch to all supported branches.
Dean Rasheed, reviewed by Joel Jacobson.
Discussion: https://postgr.es/m/CAEZATCXB%2BrDTuMjhK5ZxcouufigSc-X4tGJCBTMpZ3n%3DxxQuhg%40mail.gmail.com
ProcessIncomingNotify is called from the main processing loop that
normally runs in MessageContext. That outer-loop code assumes that
whatever it allocates will be cleaned up when we're done processing
the current client message --- but if we service a notify interrupt,
then whatever gets allocated before the next switch into
MessageContext will be permanently leaked in TopMemoryContext,
because CommitTransactionCommand sets CurrentMemoryContext to
TopMemoryContext. There are observable leaks associated with
(at least) encoding conversion of incoming queries and parameters
attached to Bind messages.
sinval catchup interrupts have a similar problem. There might be
others, but I've not identified any other clear cases.
To fix, take care to save and restore CurrentMemoryContext across
the Start/CommitTransactionCommand calls in these functions.
Per bug #18512 from wizardbrony. Commit to back branches only;
in HEAD, this was dealt with by the riskier but more thoroughgoing
approach in commit 1afe31f03.
Discussion: https://postgr.es/m/3478884.1718656625@sss.pgh.pa.us
This extends ad98fb1422 to invals of
inplace updates. Trouble requires an inplace update of a catalog having
a TOAST table, so only pg_database was at risk. (The other catalog on
which core code performs inplace updates, pg_class, has no TOAST table.)
Trouble would require something like the inplace-inval.spec test.
Consider GRANT ... ON DATABASE fetching a stale row from cache and
discarding a datfrozenxid update that vac_truncate_clog() has already
relied upon. Back-patch to v12 (all supported versions).
Reviewed (in an earlier version) by Robert Haas.
Discussion: https://postgr.es/m/20240114201411.d0@rfd.leadboat.com
Discussion: https://postgr.es/m/20240512232923.aa.nmisch@google.com
This has no user-visible, important consequences, since other sessions'
catalog scans can't find the relation until we commit. However, this
unblocks introducing a rule about locks required to heap_update() a
pg_class row. CREATE TABLE has been acquiring this lock eventually, but
it can heap_update() pg_class.relchecks earlier. create_toast_table()
has been acquiring only ShareLock. Back-patch to v12 (all supported
versions), the plan for the commit relying on the new rule.
Reviewed (in an earlier version) by Robert Haas.
Discussion: https://postgr.es/m/20240611024525.9f.nmisch@google.com
Commit 5b562644fe added a comment that
SetRelationHasSubclass() callers must hold this lock. When commit
17f206fbc8 extended use of this column to
partitioned indexes, it didn't take the lock. As the latter commit
message mentioned, we currently never reset a partitioned index to
relhassubclass=f. That largely avoids harm from the lock omission. The
cause for fixing this now is to unblock introducing a rule about locks
required to heap_update() a pg_class row. This might cause more
deadlocks. It gives minor user-visible benefits:
- If an ALTER INDEX SET TABLESPACE runs concurrently with ALTER TABLE
ATTACH PARTITION or CREATE PARTITION OF, one transaction blocks
instead of failing with "tuple concurrently updated". (Many cases of
DDL concurrency still fail that way.)
- Match ALTER INDEX ATTACH PARTITION in choosing to lock the index.
While not user-visible today, we'll need this if we ever make something
set the flag to false for a partitioned index, like ANALYZE does today
for tables. Back-patch to v12 (all supported versions), the plan for
the commit relying on the new rule. In back branches, add
LockOrStrongerHeldByMe() instead of adding a LockHeldByMe() parameter.
Reviewed (in an earlier version) by Robert Haas.
Discussion: https://postgr.es/m/20240611024525.9f.nmisch@google.com
These commands already make the persistence of owned sequences follow
owned table persistence changes. They didn't lock those sequences.
They lost the effect of nextval() calls that other sessions make after
the ALTER TABLE command, before the ALTER TABLE transaction commits.
Fix by acquiring the same lock that ALTER SEQUENCE SET { LOGGED |
UNLOGGED } acquires. This might cause more deadlocks. Back-patch to
v15, where commit 344d62fb9a introduced
unlogged sequences.
Reviewed (in an earlier version) by Robert Haas.
Discussion: https://postgr.es/m/20240611024525.9f.nmisch@google.com
Most comments concern RELKIND_VIEW. One addresses the ExecUpdate()
"tupleid" parameter. A later commit will rely on these facts, but they
hold already. Back-patch to v12 (all supported versions), the plan for
that commit.
Reviewed (in an earlier version) by Robert Haas.
Discussion: https://postgr.es/m/20240512232923.aa.nmisch@google.com
errfinish() assumes that the __FUNC__ and __FILE__ arguments it's
passed are compile-time constant strings that can just be pointed
to rather than physically copied. However, it's possible for LLVM
to generate code in which those pointers point into a dynamically
loaded code segment. If that segment gets unloaded before we're
done with the ErrorData struct, we have dangling pointers that
will lead to SIGSEGV. In simple cases that won't happen, because we
won't unload LLVM code before end of transaction. But it's possible
to happen if the error is thrown within end-of-transaction code run by
_SPI_commit or _SPI_rollback, because since commit 2e517818f those
functions clean up by ending the transaction and starting a new one.
Rather than fixing this by adding pstrdup() overhead to every
elog/ereport sequence, let's fix it by copying the risky pointers
in CopyErrorData(). That solves it for _SPI_commit/_SPI_rollback
because they use that function to preserve the error data across
the transaction end/restart sequence; and it seems likely that
any other code doing something similar would need to do that too.
I'm suspicious that this behavior amounts to an LLVM bug (or a
bug in our use of it?), because it implies that string constant
references that should be pointer-equal according to a naive
understanding of C semantics will sometimes not be equal.
However, even if it is a bug and someday gets fixed, we'll have
to cope with the current behavior for a long time to come.
Report and patch by me. Back-patch to all supported branches.
Discussion: https://postgr.es/m/1565654.1719425368@sss.pgh.pa.us
We did not recover the subtransaction IDs of prepared transactions
when starting a hot standby from a shutdown checkpoint. As a result,
such subtransactions were considered as aborted, rather than
in-progress. That would lead to hint bits being set incorrectly, and
the subtransactions suddenly becoming visible to old snapshots when
the prepared transaction was committed.
To fix, update pg_subtrans with prepared transactions's subxids when
starting hot standby from a shutdown checkpoint. The snapshots taken
from that state need to be marked as "suboverflowed", so that we also
check the pg_subtrans.
Backport to all supported versions.
Discussion: https://www.postgresql.org/message-id/6b852e98-2d49-4ca1-9e95-db419a2696e0@iki.fi
In pgoutput, when converting the child table's tuple format to match the
parent table's, we temporarily create a new slot to store the converted
tuple. However, we missed to drop such temporary slots, leading to
resource leakage.
Reported-by: Bowen Shi
Author: Hou Zhijie
Reviewed-by: Amit Kapila
Backpatch-through: 15
Discussion: https://postgr.es/m/CAM_vCudv8dc3sjWiPkXx5F2b27UV7_YRKRbtSCcE-pv=cVACGA@mail.gmail.com
When pgstats is initialized for a backend, it uses dsa_attach_in_place()
without a "segment" provided. Hence, no callback is registered to
automatically release the DSA attached once a backend exits. Not doing
any cleanup causes the reference count of the pgstats DSA to
continuously increment, at some point overflowing it (the more the
number of connections, the faster it is to reach this state). Once the
reference count overflows and then gets back to 0, new backends are not
able to attach to the pgstats DSA, failing startup.
This issue is resolved by adding in the pgstats shutdown hook a call to
dsa_release_in_place(), ensuring that the DSA attached at backend
startup is correctly released, keeping the reference count at bay.
The author of this patch has been able to see this issue on a server
with a long uptime and a high connection turnover.
Issue introduced by 5891c7a8ed, so backpatch down to 15.
Author: Anthonin Bonnefoy
Discussion: https://postgr.es/m/CAO6_XqqJbJBL=M7Ym13TcB4Xnq58vRa2jcC+gwEPBgbAda6B1Q@mail.gmail.com
Backpatch-through: 15
1. TruncateMultiXact() performs the SLRU truncations in a critical
section. Deleting the SLRU segments calls ForwardSyncRequest(), which
will try to compact the request queue if it's full
(CompactCheckpointerRequestQueue()). That in turn allocates memory,
which is not allowed in a critical section. Backtrace:
TRAP: failed Assert("CritSectionCount == 0 || (context)->allowInCritSection"), File: "../src/backend/utils/mmgr/mcxt.c", Line: 1353, PID: 920981
postgres: autovacuum worker template0(ExceptionalCondition+0x6e)[0x560a501e866e]
postgres: autovacuum worker template0(+0x5dce3d)[0x560a50217e3d]
postgres: autovacuum worker template0(ForwardSyncRequest+0x8e)[0x560a4ffec95e]
postgres: autovacuum worker template0(RegisterSyncRequest+0x2b)[0x560a50091eeb]
postgres: autovacuum worker template0(+0x187b0a)[0x560a4fdc2b0a]
postgres: autovacuum worker template0(SlruDeleteSegment+0x101)[0x560a4fdc2ab1]
postgres: autovacuum worker template0(TruncateMultiXact+0x2fb)[0x560a4fdbde1b]
postgres: autovacuum worker template0(vac_update_datfrozenxid+0x4b3)[0x560a4febd2f3]
postgres: autovacuum worker template0(+0x3adf66)[0x560a4ffe8f66]
postgres: autovacuum worker template0(AutoVacWorkerMain+0x3ed)[0x560a4ffe7c2d]
postgres: autovacuum worker template0(+0x3b1ead)[0x560a4ffecead]
postgres: autovacuum worker template0(+0x3b620e)[0x560a4fff120e]
postgres: autovacuum worker template0(+0x3b3fbb)[0x560a4ffeefbb]
postgres: autovacuum worker template0(+0x2f724e)[0x560a4ff3224e]
/lib/x86_64-linux-gnu/libc.so.6(+0x27c8a)[0x7f62cc642c8a]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0x85)[0x7f62cc642d45]
postgres: autovacuum worker template0(_start+0x21)[0x560a4fd16f31]
To fix, bail out in CompactCheckpointerRequestQueue() without doing
anything, if it's called in a critical section. That covers the above
call path, as well as any other similar cases where
RegisterSyncRequest might be called in a critical section.
2. After fixing that, another problem became apparent: Autovacuum
process doing that truncation can deadlock with the checkpointer
process. TruncateMultiXact() sets "MyProc->delayChkptFlags |=
DELAY_CHKPT_START". If the sync request queue is full and cannot be
compacted, the process will repeatedly sleep and retry, until there is
room in the queue. However, if the checkpointer is trying to start a
checkpoint at the same time, and is waiting for the DELAY_CHKPT_START
processes to finish, the queue will never shrink.
More concretely, the autovacuum process is stuck here:
#0 0x00007fc934926dc3 in epoll_wait () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x000056220b24348b in WaitEventSetWaitBlock (set=0x56220c2e4b50, occurred_events=0x7ffe7856d040, nevents=1, cur_timeout=<optimized out>) at ../src/backend/storage/ipc/latch.c:1570
#2 WaitEventSetWait (set=0x56220c2e4b50, timeout=timeout@entry=10, occurred_events=<optimized out>, occurred_events@entry=0x7ffe7856d040, nevents=nevents@entry=1,
wait_event_info=wait_event_info@entry=150994949) at ../src/backend/storage/ipc/latch.c:1516
#3 0x000056220b243224 in WaitLatch (latch=<optimized out>, latch@entry=0x0, wakeEvents=wakeEvents@entry=40, timeout=timeout@entry=10, wait_event_info=wait_event_info@entry=150994949)
at ../src/backend/storage/ipc/latch.c:538
#4 0x000056220b26cf46 in RegisterSyncRequest (ftag=ftag@entry=0x7ffe7856d0a0, type=type@entry=SYNC_FORGET_REQUEST, retryOnError=true) at ../src/backend/storage/sync/sync.c:614
#5 0x000056220af9db0a in SlruInternalDeleteSegment (ctl=ctl@entry=0x56220b7beb60 <MultiXactMemberCtlData>, segno=segno@entry=11350) at ../src/backend/access/transam/slru.c:1495
#6 0x000056220af9dab1 in SlruDeleteSegment (ctl=ctl@entry=0x56220b7beb60 <MultiXactMemberCtlData>, segno=segno@entry=11350) at ../src/backend/access/transam/slru.c:1566
#7 0x000056220af98e1b in PerformMembersTruncation (oldestOffset=<optimized out>, newOldestOffset=<optimized out>) at ../src/backend/access/transam/multixact.c:3006
#8 TruncateMultiXact (newOldestMulti=newOldestMulti@entry=3221225472, newOldestMultiDB=newOldestMultiDB@entry=4) at ../src/backend/access/transam/multixact.c:3201
#9 0x000056220b098303 in vac_truncate_clog (frozenXID=749, minMulti=<optimized out>, lastSaneFrozenXid=749, lastSaneMinMulti=3221225472) at ../src/backend/commands/vacuum.c:1917
#10 vac_update_datfrozenxid () at ../src/backend/commands/vacuum.c:1760
#11 0x000056220b1c3f76 in do_autovacuum () at ../src/backend/postmaster/autovacuum.c:2550
#12 0x000056220b1c2c3d in AutoVacWorkerMain (startup_data=<optimized out>, startup_data_len=<optimized out>) at ../src/backend/postmaster/autovacuum.c:1569
and the checkpointer is stuck here:
#0 0x00007fc9348ebf93 in clock_nanosleep () from /lib/x86_64-linux-gnu/libc.so.6
#1 0x00007fc9348fe353 in nanosleep () from /lib/x86_64-linux-gnu/libc.so.6
#2 0x000056220b40ecb4 in pg_usleep (microsec=microsec@entry=10000) at ../src/port/pgsleep.c:50
#3 0x000056220afb43c3 in CreateCheckPoint (flags=flags@entry=108) at ../src/backend/access/transam/xlog.c:7098
#4 0x000056220b1c6e86 in CheckpointerMain (startup_data=<optimized out>, startup_data_len=<optimized out>) at ../src/backend/postmaster/checkpointer.c:464
To fix, add AbsorbSyncRequests() to the loops where the checkpointer
waits for DELAY_CHKPT_START or DELAY_CHKPT_COMPLETE operations to
finish.
Backpatch to v14. Before that, SLRU deletion didn't call
RegisterSyncRequest, which avoided this failure. I'm not sure if there
are other similar scenarios on older versions, but we haven't had
any such reports.
Discussion: https://www.postgresql.org/message-id/ccc66933-31c1-4f6a-bf4b-45fef0d4f22e@iki.fi
When detaching partition in concurrent mode, it's possible for partition
descriptors to not match the set that was recently seen when the plan
was made, causing an assertion failure or (in production builds) failure
to construct a working plan. The case that was reported involves
prepared statements, but I think it may be possible to hit this bug
without that too.
The problem is that CreatePartitionPruneState is constructing a
PartitionPruneState under the assumption that new partitions can be
added, but never removed, but it turns out that this isn't true: a
prepared statement gets replanned when the DETACH CONCURRENTLY session
sends out its invalidation message, but if the invalidation message
arrives after ExecInitAppend started, we would build a partition
descriptor without the partition, and then CreatePartitionPruneState
would refuse to work with it.
CreatePartitionPruneState already contains code to deal with the new
descriptor having more partitions than before (and behaving for the
extra partitions as if they had been pruned), but doesn't have code to
deal with less partitions than before, and it is naïve about the case
where the number of partitions is the same. We could simply add that a
new stanza for less partitions than before, and in simple testing it
works to do that; but it's possible to press the test scripts even
further and hit the case where one partition is added and a partition is
removed quickly enough that we see the same number of partitions, but
they don't actually match, causing hangs during execution.
To cope with both these problems, we now memcmp() the arrays of
partition OIDs, and do a more elaborate mapping (relying on the fact
that both OID arrays are in partition-bounds order) if they're not
identical.
Backpatch to 14, where DETACH CONCURRENTLY appeared.
Reported-by: yajun Hu <1026592243@qq.com>
Reviewed-by: Tender Wang <tndrwang@gmail.com>
Discussion: https://postgr.es/m/18377-e0324601cfebdfe5@postgresql.org
afterTriggerInvokeEvents and AfterTriggerExecute have always
treated it as an error if the trigger OID mentioned in a queued
after-trigger event can't be found. However, that fails to
account for the edge case where the trigger's been dropped in
the current transaction since queueing the event. There seems
no very good reason to disallow that case, so instead silently
do nothing if the trigger OID can't be found.
This does give up a little bit of bug-detection ability, but I don't
recall that these error messages have ever actually revealed a bug,
so it seems mostly theoretical. Alternatives such as marking
pending events DONE at the time of dropping a trigger would be
complicated and perhaps introduce bugs of their own.
Per bug #18517 from Alexander Lakhin. Back-patch to all
supported branches.
Discussion: https://postgr.es/m/18517-af2d19882240902c@postgresql.org
In cost_memoize_rescan(), when calculating the hit_ratio using the calls
and ndistinct estimations, if the value that was set in
MemoizePath.calls had not been processed through clamp_row_est(), then it
was possible that it was set to some non-integer value which could result
in ndistinct being 1 higher than calls due to estimate_num_groups()
performing clamp_row_est() on its input_rows. This could result in
hit_ratio values slightly below 0.0, which would cause an Assert failure.
The value of MemoizePath.calls comes from the final parameter in the
create_memoize_path() function, of which we only have one true caller of.
That caller passes outer_path->rows. All the core code I looked at
always seems to call clamp_row_est() on the Path.rows, so there might
have been no issues with any core Paths causing troubles here. The bug
report was about a CustomPath with a non-clamped row estimated.
The misbehavior as a result of this seems to be mostly limited to the
Assert() failing. Aside from that, it seems the Memoize costs would
just come out slightly higher than they should have, which is likely
fairly harmless.
Reported-by: Kohei KaiGai <kaigai@heterodb.com>
Discussion: https://postgr.es/m/CAOP8fzZnTU+N64UYJYogb1hN-5hFP+PwTb3m_cnGAD7EsQwrKw@mail.gmail.com
Reviewed-by: Richard Guo
Backpatch-through: 14, where Memoize was introduced
Reconstruction of an SP-GiST index by REINDEX CONCURRENTLY may
insert some REDIRECT tuples. This will typically happen in
a transaction that lacks an XID, which leads either to assertion
failure in spgFormDeadTuple or to insertion of a REDIRECT tuple
with zero xid. The latter's not good either, since eventually
VACUUM will apply GlobalVisTestIsRemovableXid() to the zero xid,
resulting in either an assertion failure or a garbage answer.
In practice, since REINDEX CONCURRENTLY locks out index scans
till it's done, it doesn't matter whether it inserts REDIRECTs
or PLACEHOLDERs; and likewise it doesn't matter how soon VACUUM
reduces such a REDIRECT to a PLACEHOLDER. So in non-assert builds
there's no observable problem here, other than perhaps a little
index bloat. But it's not behaving as intended.
To fix, remove the failing Assert in spgFormDeadTuple, acknowledging
that we might sometimes insert a zero XID; and guard VACUUM's
GlobalVisTestIsRemovableXid() call with a test for valid XID,
ensuring that we'll reduce such a REDIRECT the first time VACUUM
sees it. (Versions before v14 use TransactionIdPrecedes here,
which won't fail on zero xid, so they really have no bug at all
in non-assert builds.)
Another solution could be to not create REDIRECTs at all during
REINDEX CONCURRENTLY, making the relevant code paths treat that
case like index build (which likewise knows that no concurrent
index scans can be happening). That would allow restoring the
Assert in spgFormDeadTuple, but we'd still need the VACUUM change
because redirection tuples with zero xid may be out there already.
But there doesn't seem to be a nice way for spginsert() to tell that
it's being called in REINDEX CONCURRENTLY without some API changes,
so we'll leave that as a possible future improvement.
In HEAD, also rename the SpGistState.myXid field to redirectXid,
which seems less misleading (since it might not in fact be our
transaction's XID) and is certainly less uninformatively generic.
Per bug #18499 from Alexander Lakhin. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18499-8a519c280f956480@postgresql.org
DeleteInitPrivs did not get the memo about how, when dropping a
whole object (with subid == 0), you should drop entries relating
to its sub-objects too. This is visible in the test_pg_dump test
case if one drops the extension at the end: the entry for
GRANT SELECT(col1) ON regress_pg_dump_table TO public;
was still present in pg_init_privs afterwards, although it was
pointing to a dangling table OID.
Noted while fooling with a fix for REASSIGN OWNED for pg_init_privs
entries. This bug is aboriginal in the pg_init_privs feature
though, and there seems no reason not to back-patch the fix.
The manual says clearly that punctuation in the input of
websearch_to_tsquery() is ignored, except for the special cases
of dashes and quotes. However, this failed for cases like
"(foo bar) or something", or in general an ISOPERATOR character
in front of the "or". We'd switch back to WAITOPERAND state,
then ignore the operator character while remaining in that state,
and then reach the "or" in WAITOPERAND state which (intentionally)
makes us treat it as data.
The fix is simple enough: if we see an ISOPERATOR character while in
WAITOPERATOR state, we have to skip it while staying in that state.
(We don't need to worry about other punctuation characters: those will
be consumed as though they were words, but then rejected by lexizing.)
In v14 and up (since commit eb086056f) we can simplify the code a bit
more too, because there is no longer a reason for the WAITOPERAND
state to distinguish between quoted and unquoted operands.
Per bug #18479 from Manos Emmanouilidis. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18479-d9b46e2fc242c33e@postgresql.org
The purpose of the function is to reduce the effective
autovacuum_multixact_freeze_max_age if the multixact members SLRU is
approaching wraparound, to make multixid freezing more aggressive.
The returned value should therefore never be greater than plain
autovacuum_multixact_freeze_max_age.
Reviewed-by: Robert Haas
Discussion: https://www.postgresql.org/message-id/85fb354c-f89f-4d47-b3a2-3cbd461c90a3@iki.fi
Backpatch-through: 12, all supported versions
infer_arbiter_indexes failed to renumber varnos in index expressions
or predicates that it got from the catalogs. This escaped detection
up to now because the stored varnos in such trees will be 1, and an
INSERT's result relation is usually the first rangetable entry,
so that that was fine. However, in cases such as inserting through
an updatable view, it's not fine, leading to failure to match the
expressions to the query with ensuing "there is no unique or exclusion
constraint matching the ON CONFLICT specification" errors.
Fix by copy-and-paste from get_relation_info().
Per bug #18502 from Michael Wang. Back-patch to all supported
versions.
Discussion: https://postgr.es/m/18502-545b53f5b81e54e0@postgresql.org
When a partition is being detached in concurrent mode, it is possible
for find_inheritance_children_extended() to return that partition in the
list, and immediately after that receive an invalidation message that
sets its relpartbound to NULL just before we read it. (This can happen
because table_open() reads invalidation messages.) Currently we raise
an error
ERROR: missing relpartbound for relation %u
about the situation, but that's bogus because the table is no longer a
partition, so we shouldn't be complaining about it. A better reaction
is to retry the find_inheritance_children_extended call to get a new
list, which will no longer have the partition being detached.
Noticed while investigating bug #18377.
Backpatch to 14, where DETACH CONCURRENTLY appeared.
Discussion: https://postgr.es/m/202405201616.y4ht2qe5ihoy@alvherre.pgsql
Normally this case isn't even reachable by non-superusers, since
permissions checks prevent naming such a table. However, it is
possible to make it happen by altering a parent table whose child
is another session's temp table.
We definitely can't support any such ALTER that requires modifying
the contents of such a table, since we lack access to the other
session's temporary-buffer pool. But there seems no good reason
to allow it even if it'd only require changing catalog contents.
One reason not to allow it is that we'd rather not expose the
implementation-dependent behavior of whether a specific ALTER
requires touching the table contents. Another is that there may
be (in future, even if not today) optimizations that assume that
a session's own temp tables won't be modified by other sessions.
Hence, add a RELATION_IS_OTHER_TEMP() check to all the places
where ALTER TABLE currently does CheckTableNotInUse(). (I looked
through all other callers of CheckTableNotInUse(), and they seem
OK already.)
Per bug #18492 from Alexander Lakhin. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18492-c7a2634bf4968763@postgresql.org
If the CALL is within an atomic context (e.g. there's an outer
transaction block), _SPI_execute_plan should acquire a fresh snapshot
to execute any such functions with. We failed to do that and instead
passed them the Portal snapshot, which had been acquired at the start
of the current SQL command. This'd lead to seeing stale values of
rows modified since the start of the command.
This is arguably a bug in 84f5c2908: I failed to see that "are we in
non-atomic mode" needs to be defined the same way as it is further
down in _SPI_execute_plan, i.e. check !_SPI_current->atomic not just
options->allow_nonatomic. Alternatively the blame could be laid on
plpgsql, which is unconditionally passing allow_nonatomic = true
for CALL/DO even when it knows it's in an atomic context. However,
fixing it in spi.c seems like a better idea since that will also fix
the problem for any extensions that may have copied plpgsql's coding
pattern.
While here, update an obsolete comment about _SPI_execute_plan's
snapshot management.
Per report from Victor Yegorov. Back-patch to all supported versions.
Discussion: https://postgr.es/m/CAGnEboiRe+fG2QxuBO2390F7P8e2MQ6UyBjZSL_w1Cej+E4=Vw@mail.gmail.com
Floris Van Nee has reported a bug in the pgstats facility where a stats
entry already dropped would get again dropped. This case should not
happen, still the error generated did not offer any details about the
stats entry getting dropped.
This commit improves the error message generated to inform about the
stats entry kind, database OID, object OID and refcount, which should
help to debug more the problem reported. Bertrand Drouvot has been
independently able to reach this error path while writing a new feature,
and more details about the failure would have been helpful for
debugging.
Author: Andres Freund, Bertrand Drouvot
Discussion: https://postgr.es/m/20240505160915.6boysum4f34siqct@awork3.anarazel.de
Discussion: https://postgr.es/m/ZkM30paAD8Cr/Bix@ip-10-97-1-34.eu-west-3.compute.internal
Backpatch-through: 15
Concurrent activity around replication slot creation and drop could
cause a replication slot to use a stats entry it should not have used
when created, triggering an assertion failure when retrieving this
inconsistent entry from the dshash table used by the stats facility.
The issue is that pgstat_drop_replslot() calls pgstat_drop_entry()
without checking the result. If pgstat_drop_entry() cannot free the
entry related to the object dropped, pgstat_request_entry_refs_gc()
should be called. AtEOXact_PgStat_DroppedStats() and surrounding
routines dropping stats entries already do that.
This is documented in pgstat_internal.h, but let's add a comment at the
top of pgstat_drop_entry() as that can be easy to miss.
Reported-by: Alexander Lakhin
Author: Floris Van Nee
Analyzed-by: Andres Freund
Discussion: https://postgr.es/m/17947-b9554521ad963c9c@postgresql.org
Backpatch-through: 15
transformTableLikeClause believed that it could process extended
statistics immediately because "the representation of CreateStatsStmt
doesn't depend on column numbers". That was true when extended stats
were first introduced, but it was falsified by the addition of
extended stats on expressions: the parsed expression tree is fed
forward by the LIKE option, and that will contain Vars. So if the
new table doesn't have attnums identical to the old one's (typically
because there are some dropped columns in the old one), that doesn't
work. The CREATE goes through, but it emits invalid statistics
objects that will cause problems later.
Fortunately, we already have logic that can adapt expression trees
to the possibly-new column numbering. To use it, we have to delay
processing of CREATE_TABLE_LIKE_STATISTICS into expandTableLikeClause,
just as for other LIKE options that involve expressions.
Per bug #18468 from Alexander Lakhin. Back-patch to v14 where
extended statistics on expressions were added.
Discussion: https://postgr.es/m/18468-f5add190e3fa5902@postgresql.org
We are capable of optimizing MIN() and MAX() aggregates on indexed
columns into subqueries that exploit the index, rather than the normal
thing of scanning the whole table. When we do this, we replace the
Aggref node(s) with Params referencing subquery outputs. Such Params
really ought to be included in the per-plan-node extParam/allParam
sets computed by SS_finalize_plan. However, we've never done so
up to now because of an ancient implementation choice to perform
that substitution during set_plan_references, which runs after
SS_finalize_plan, so that SS_finalize_plan never sees these Params.
The cleanest fix would be to perform a separate tree walk to do
these substitutions before SS_finalize_plan runs. That seems
unattractive, first because a whole-tree mutation pass is expensive,
and second because we lack infrastructure for visiting expression
subtrees in a Plan tree, so that we'd need a new function knowing
as much as SS_finalize_plan knows about that. I also considered
swapping the order of SS_finalize_plan and set_plan_references,
but that fell foul of various assumptions that seem tricky to fix.
So the approach adopted here is to teach SS_finalize_plan itself
to check for such Aggrefs. I refactored things a bit in setrefs.c
to avoid having three copies of the code that does that.
Back-patch of v17 commits d0d44049d and 779ac2c74. When d0d44049d
went in, there was no evidence that it was fixing a reachable bug,
so I refrained from back-patching. Now we have such evidence.
Per bug #18465 from Hal Takahara. Back-patch to all supported
branches.
Discussion: https://postgr.es/m/18465-2fae927718976b22@postgresql.org
Discussion: https://postgr.es/m/2391880.1689025003@sss.pgh.pa.us
Specifically, it terminates a background worker even if the caller
couldn't terminate the background worker with pg_terminate_backend().
Commit 3a9b18b309 neglected to update
this. Back-patch to v13, which introduced DROP DATABASE FORCE.
Reviewed by Amit Kapila. Reported by Kirill Reshke.
Discussion: https://postgr.es/m/20240429212756.60.nmisch@google.com
Most of the infrastructure for procedure arguments was already
okay with polymorphic output arguments, but it turns out that
CallStmtResultDesc() was a few bricks shy of a load here. It thought
all it needed to do was call build_function_result_tupdesc_t, but
that function specifically disclaims responsibility for resolving
polymorphic arguments. Failing to handle that doesn't seem to be
a problem for CALL in plpgsql, but CALL from plain SQL would get
errors like "cannot display a value of type anyelement", or even
crash outright.
In v14 and later we can simply examine the exposed types of the
CallStmt.outargs nodes to get the right type OIDs. But it's a lot
more complicated to fix in v12/v13, because those versions don't
have CallStmt.outargs, nor do they do expand_function_arguments
until ExecuteCallStmt runs. We have to duplicatively run
expand_function_arguments, and then re-determine which elements
of the args list are output arguments.
Per bug #18463 from Drew Kimball. Back-patch to all supported
versions, since it's busted in all of them.
Discussion: https://postgr.es/m/18463-f8cd77e12564d8a2@postgresql.org
