PostgreSQL 15 Open Items

Open Issues

NOTE: Please place new open items at the end of the list.

Older bugs affecting stable branches

Live issues

• RecoveryConflictInterrupt() is unsafe in a signal handler
  • This seems to explain buildfarm failures in 031_recovery_conflict.pl
  • Affects all stable branches.

• REINDEX on a system catalog can leave index with two index tuples whose heap TIDs match
  • In other words, there is a rare case where the HOT invariant is violated. Same HOT chain is indexed twice due to confusion about which precise heap tuple should be indexed.
  • Unclear what the user impact is.
  • Affects all stable branches.

• memory leak with JIT inlining
  • Another report of (same?) issue and reproducer
  • Another report
  • Another report

• Non-replayable WAL records through overflows and >MaxAllocSize lengths
  • In other words; we can write xlog records that we can't read (plus potentially actual WAL corruption); making the instance unrecoverable, and blocks any replication.
  • Exploitation seems limited to WAL records of 2PC and logical replication, and extension-generated WAL.
  • Affects all stable branches.

• TOAST fetches could perhaps occur after the needed data has been removed
  • The symptom originally reported in the thread was fixed by 9f4f0a0dad4c7422a97d94e4051c08ec6d181dd6, but nobody is very happy with the status quo in this area. Do we need to do more now?
  • Affects all stable branches.

Fixed issues

• CLUSTER sort on abbreviated expressions is broken
  • Affects all stable branches.
  • Fixed by: 8ab0ebb9a842dc6063d1374a38b47a3b7ee64afe

• Records missing from Primary Key index when doing REINDEX INDEX CONCURRENTLY
  • Affects v14
  • Fixed by: e28bb885196916b0a3d898ae4f2be0e38108d81b

• -c min_dynamic_shared_memory now triggers an assertion
  • Affects v14
  • Fixed by: 7201cd1862

• Extension pg_trgm, permissions and pg_dump order
  • Affects all stable branches.
  • Fixed by 00377b9a02b89a831ae50e1c718d34565356698f

Resolved Issues

resolved before 15.0

• pgstat: START_REPLICATION SLOT assertion failure
  • Owner: Andres Freund
  • Fixed by: 06dbd619bfbfe03fefa7223838690d4012f874ad

resolved before 15rc2

• Memory leak in single-column sorts
  • Fixed by: f7ae8a2e186ca27f4ea64fddc21bfabaf0618507
• PQsslAttribute not behaving as documented
  • Fixed by: cae4688ce81b8449aa6e1e7bfa384d53520a81fb
• GROUP BY optimization patch is too broken to ship
  • Owner: Tomas Vondra (db0d67db2)
  • Fixed by: 443df6e2db932a7cd6d85ddfb67e11a43345130d
• psql \g misbehaves with multiple commands
  • Fixed by: 595580aa1243df5f199516ff11a27ba9680e6904

resolved before 15rc1

• recovery fails with recovery_prefetch=on and maintenance_io_concurency=0
  • Fixed by: adb466150b44d1eaf43a2d22f58ff4c545a0ed3f
• FailedAssertion("TransactionIdIsValid(xmax)
  • Fixed by: adb466150b44d1eaf43a2d22f58ff4c545a0ed3f
• pg_publication_tables view is busted
  • Fixed by: 20b6847176976da9cae6103475b967e6c2971898
• replaying wal_compression=lz4/zstd on a server which doesn't support it exposes an elog() which includes no information about the failure
  • Owner: Michael Paquier (4035cd5)
  • Fixed by: df4a056619a7e9868ec43a15e2088ff8ccf94471
• syntax error in MERGE docs
  • Fixed by: 5dcebd10baa8ec14a97029ca0bacaf2a9a0f8652
• pg_upgrade generated files in subdir follow-up
  • Fixed by: ee5353abb6124de5ffd24ef1cedbc2a7196d4fd5
• pg_basebackup's --gzip switch misbehaves
  • Fixed by: f352e2d08ac048d7407dd6098fc6b344ff85c2dd
• Assertion failure in WaitForWALToBecomeAvailable state machine
  • Fixed by: b4f584f9d2a1b50d1f5ce88408a54bb11f1c8353
• Possible to create database with incompatible ICU and encoding settings
  • Peter Eisentraut (f2553d43060edb210b36c63187d52a632448e1d2)
  • Fixed by: c7db01e325a530ec38ec7ba57cd3ed32e123e33c
• startup process says "recovering NNN" after it's done recovering NNN
  • Fixed by: fbb5f54b67c2f35c885d07daa26bce7e2eb6b0be
• why can't a table be part of the same publication as its schema
  • Owner: Amit Kapila
  • Fixed by: 13a185f54ba594a5fb42bac3e1ecbdfd62b9041f

resolved before 15beta4

• MERGE fails if inside a CTE
  • Fixed by: 455d254d22665eb
• Materialized view rewrite is broken when there is an event trigger
  • Owner: Michael Paquier (b0483263dda0824cc49e3f8a022dab07e1cdf9a7)
  • Fixed by: 93f2349c36a7555a81d5e26edf9a0213fb3d323a
• CREATE DATABASE STRATEGY WAL_LOG crash and memory corruption
  • Owner: Robert Haas (9c08aea6a3090a396be334cc58c511edab05776a)
• crash in parallel vacuum
  • Fixed by: 662ba729a67e5c6e43efbc5263e1d2c3a72a8741
• Crash in ICU for global collation
  • Peter Eisentraut (f2553d43060edb210b36c63187d52a632448e1d2)
  • One issue has been fixed in 36f729e2bc3539e777cda698dc441b8ccec42142 but we still have a problem with pg_collation.colliculocale not getting dumped
  • Fixed by: 396d348b046c6b7e5dc83158c4c1df1377a1d2ef
• PANIC during standby promotion
  • Fixed by: 932b01630095a08797e05ba1dae279731b862c95
• GROUP BY optimization defeated partitionwise tests
  • Possibly-related complaint here
  • Owner: Tomas Vondra (db0d67db2)
  • Fixed by: 2fe6b2a806f20f2dbf02c6e79783dbc2777a1840

resolved before 15beta3

• tuplesort Generation memory contexts don't play nicely with index builds
  • Owner: David Rowley
  • Fixed by: ae1123f9899fe80935ae344e38f18632beb1bf9a
• Repeatability of installcheck for test_oat_hooks
  • Owner: Andrew Dunstan
  • Fixed by: a6434b951558baad8372dc4b83bf87606dac9cda
• Revert debugging added due to 019_replslot_limit
  • Owner: Andres Freund
  • Reverted: 3f8148c256e067dc2e8929ed174671ba7dc3339c
• PG15 beta1 sort performance regression due to Generation context change
  • Owner: David Rowley
• pg_parameter_aclcheck() and trusted extensions
  • Owner: Tom Lane (a0ffa885e478f5eeacc4e250e35ce25a4740c487)
  • Fixed by: 13d83881514856353dc86575eb0fc28132349a60
• Unprivileged user can induce crash by using an SUSET param in PGOPTIONS
  • Owner: Tom Lane (a0ffa885e478f5eeacc4e250e35ce25a4740c487)
  • Fixed by: b35617de37870756bdb0e00ffc0a42441e56eefa
• wrong filename used in pg_ident_file_mapping infrastructure
  • Owner: Michael Paquier (a2c84990bea7beadb599d02328190e2a763dcb86)
  • Fixed by: 27e0ee57f68d27af68967759a2ff61a581f501dc
• Endless loop with UNIQUE NULLS NOT DISTINCT and INSERT ... ON CONFLICT
  • Owner: Peter Eisentraut (94aa7cc5f707712f592885995a28e018c7c80488)
  • Fixed by: d59383924c580a77a2346d9b1284c8589b3d43e2
• XX000 error caused by window function run conditions
  • Owner: David Rowley
  • Fixed by: 270eb4b5d4986534f2d522ebb19f67396d13cf44

• large objects lost on upgrade
  • Owner: Robert Haas (9a974cbcba005256a19991203583a94b4f9a21a9)
  • Fixed by: bbe08b8869bd29d587f24ef18eb45c7d4d14afca

resolved before 15beta2

• pgbench --partitions=0
  • Owner; Michael Paquier (6f164e6d17616a157ea5d9e34dbb1b211c080c41)
  • Fixed by: 27f1366050c6cd8c1ea5f03b367a5a167ebf34b7
• psql now shows zero elapsed time after an error
  • Owner: Peter Eisentraut
  • Fixed by: 9520f8d92a8681e441cc863422babd544353dd39
• Regression in 15beta1 when filtering subquery including row_number window function
  • Owner: David Rowley (9d9c02ccd1aea8e9131d8f4edb21bf1687e40782)
  • Fixed by: 3e9abd2eb1b1f6863250f060290f514f30ce8044
• Safety of subtrans ID caching
  • Owner: Michael Paquier (06f5295af673df795e8e70e28c43d61c2817b6df)
  • Fixed by: b4529005fd387e863bfa9eb863629b1183c0449c
• pg_upgrade test writes to source directory
  • Owner: Michael Paquier (322becb6085cb92d3708635eea61b45776bf27b6)
  • Fixed by: 15b6d2155375dee2fcba072fffa03c1c8b44656c
• pg_upgrade is not idempotent, even with --check
  • Owner: Michael Paquier (38bfae36526636ef55daf7cf2a3282403587cb5b)
  • Fixed by: 4fff78f00910af0137f9de7532f8eb21d08ab1c3
• bogus: logical replication rows/cols combinations
  • Owner: Amit Kapila
  • Fixed by: fd0b9dcebda7b931a41ce5c8e86d13f2efd0af2e
• Ignoring BRIN for HOT udpates seems broken
  • Owner: Tomas Vondra (5753d4ee320b)
  • Fixed by: e3fcca0d0d2414f3a50d6fd40eddf48b7df81475
• psql no longer reports NOTICE messages promptly
  • Owner: Peter Eisentraut (7844c9918)
  • Fixed by: e77de23fbb0f4ef27090c144edcfa889bb2a06d5
• amcheck is using a wrong macro to check compressed-ness
  • Owner: Robert Haas (bd807be6935929bdefe74d1258ca08048f0aafa3)
  • Fixed by: e243de03fb4583dd4a9f0afb41493727d7946c02
• COPY WITH (HEADER MATCH) broken with custom attribute list
  • Owner: Peter Eisentraut (072132f04e55c1c3b0f1a582318da78de7334379)
  • Fixed by: ca7a0d1d368216e89359c63531a4df0b99a437e4
• PG15 beta1 fix pg_stats_ext/pg_stats_ext_exprs view manual
  • Tomas Vondra
  • Fixed by: 401f623c7b14890011b9bb9dda7639b1de4d40ad
• Incorrect version check for datlocprovider in pg_upgrade
  • Owner: Peter Eisentraut (f2553d43060edb210b36c63187d52a632448e1d2)
  • Fixed by: fa06a34d14ea053e1e405a6ab2a1c3f1631c3a5e
• Failure in TAP tests for IP address support in SANs with LibreSSL
  • Owner: Peter Eisentraut (c1932e542863f0f646f005b3492452acc57c7e66)
  • Fixed by: 901a9d53011573e45cd7b87682f0520ef3b0fd2d

resolved before 15beta1

• libpq: duplicate error message after connection loss
  • Fixed by: 93909599cdba64c8759d646983c0a4ef93de1e50

• MERGE: ERROR: variable not found in subplan target lists
  • Fixed by: ce4f46fdc814eb1b704d81640f6d8f03625d0f53

• Buildfarm warnings
  • pg_basebackup.c:1261:35: warning: storing the address of local variable archive_filename in progress_filename [-Wdangling-pointer=]
  • new in 23a1c6578 - looks like a real error @Robert Haas
  • Fixed at: 62cb7427d1e491faf8612a82c2e3711a8cd65422

• pg_basebackup serverside compression broken with stdout and manifests
  • Fixed at: b2de45f9200d9adcac50015521574696dc464ccd

• pg_basebackup: bbstreamer_lz4.c:172: bbstreamer_lz4_compressor_content: Assertion `mystreamer->base.bbs_buffer.maxlen >= out_bound' failed.
  • basebackup LZ4 to stdout
  • Owner: Robert Haas (dab298471ff2f91f33bc25bfb73e435d3ab02148)
  • Fixed at: afb529e6772b4e2b065644a2204697eeaf6c9a96

• Incorrect reset timestamp in stats after crash recovery
  • Owner: Andres Freund (5891c7a8ed8f2d3d577e7eea34dacff12d7b6bbd)
  • Fixed at: 5cd1c40b3ce9600f129fd1fea9850e1affaf31d5

• Fixes for compression options of pg_receivewal and refactoring of backup_compression.{c,h}
  • Owner: Michael Paquier (babbbb595d2322da095a1e6703171b3f1f2815cb)
  • Fixed at: 042a923ad53dfbe39a9d5012d6c3cf3c9c338884

• Remove compatibility from pg_basebackup?
  • Fixed at: 9cd28c2e5f11dfeef64a14035b82e70acead65fd

• Timing-dependent failure in 002_archiving.pl
  • Owner: Michael Paquier (46dea2419ee7895a4eb3d048317682e6f18a17e1)
  • Fixed at: e61efafcb82c605dcc78f668685223e20d2f7ad8, 1a8b110539efe18803c1fa8aa452a2178dbad9a9

• qsort performance regression
  • Owner: John Naylor (6974924347c908335607a4a2f252213d58e21b7c)
  • Fixed at: 99c754129d787ea4ce3b34b9f4c5f5e74c45ab6a

• Small issues with CLUSTER on partitioned tables
  • Owner: Alvaro Herrera (cfdd03f45e6afc632fbe70519250ec19167d6765)
  • Fixed at: 3f19e176ae0f55a653d62e1504dbe5ad8c1006a0, 21a10368eb3fce73f146d7e48b4d81496f60d965

• asynchronous execution crash in trivial_subqueryscan()
  • Owner: Etsuro Fujita (c2bb02bc2e858ba345b8b33f1f3a54628f719d93)
  • Fixed at: 5c854e7a2c8a6cd26040e0f9949e7a4a007f6366

• Corruption due to relfilenode reuse
  • pg_upgrade can corrupt data with the new OIDs preservation feature
    • Fixed at: e2f65f42555ff531c6d7c8f151526b4ef7c016f8
  • the ProcSignalBarrier solution this builds on also turns out to have a small race/hole
    • Fixed at: b74e94dc27fdbb13954f230b1d1298430afa6c0c
  • Owner: Thomas Munro, Robert Haas

• Some issues with the TAP tests of pg_upgrade
  • Owner: Michael Paquier
  • Fixed at: 7dd3ee508432730d15c5d3032f37362f6b6e4dd8

• Crash in _outPathTarget
  • Owner: Peter Eisentraut
  • Fixed at: 9ddf251f94090cebf1bd8fc18396cb8a4b580d04

• Instability of regression test of pg_walinspect
  • Owner: Jeff Davis
  • Fixed at: ed57cac84d1c5642737dab1e4c4b8cb4f0c4305f

• crash with JSON constructors and window functions
  • Owner: Andrew Dunstan (f4fb45d15c59d7add2e1b81a9d477d0119a9691a)
  • Fixed at: 4eb9798879680dcc0e3ebb301cf6f925dfa69422, 112fdb3528465cc14a2f1dff3dc27f100326d885

• 013_partition.pl failing
  • Fixed at: dd4ab6fd6528e160571986fa8817cee9f2645aa8

• Avoid garbage logs with postgres -C on runtime-computed GUCs
  • Fixed at: 8bbf8461a3a2a38ce5f2952a025385b6938a61f7
  • Owner: Michael Paquier

• Some issues with mark_pgdllimport.pl
  • Fixed at: 5edeb574285ecbcc47f0b769a7e363404db0155b
  • Owner: Robert Haas

• Crash in new pgstats code
  • Initially reported issue was fixed by 4a736a161c306fcfed970e6b649f2f03f465ac24, but there may be more to do here.
  • Owner: Andres Freund

• TRAP: FailedAssertion("tabstat->trans == trans", File: "pgstat_relation.c", Line: 508
  • Fixed at: 0cf16cb8ca4853b084c40eca310c4c9c3ebf7e2a
  • Owner: Andres Freund

• limitations of GetMaxBackends()
  • Fixed at: 4f2400cb3f10aa79f99fba680c198237da28dd38, ab02d702ef08343fba30d90fdf7df5950063e8c9, 7fc0e7de9fb8306e84d1c15211aba4308f694455
  • Owner: Robert Haas (aa64f23b02924724eafbd9eadbf26d85df30a12b, and 4567596316d186c6e61c72df013797266fcac2f7)

Won't Fix

• InvokeNamespaceSearchHook calls need to be moved
  • Re: New Object Access Type hooks
  • Problem showed by 90efa2f5565d28054c30c18f6a2f17f94fdff91e.
• Reduce amount of logs generated by TAP tests of pg_upgrade?
  • Owner: Michael Paquier
  • Other thread: here
  • The problem is wider than just the upgrade tests, as all the runs of pg_regress would be impacted. We may want a more centralized solution for this older problem.
• Assertion failure on PG15 with modified test_shm_mq test
  • Owner: Andres Freund
  • on the thread, andres says that doesn't seem realistic to fix this, extensions will need to work around it
• PG 15 (and to a smaller degree 14) regression due to ExprEvalStep size
  • Owner: Andrew Dunstan
  • Feature reverted. Fixes will be made in a future release.

Important Dates

Current schedule:

• Feature Freeze: April 7, 2022 (Last Day to Commit Features)
• Beta 1: May 19, 2022
• Beta 2: June 30, 2022
• Beta 3: August 11, 2022
• Beta 4: September 8, 2022
• RC 1: September 29, 2022
• RC 2: October 6, 2022
• GA: October 13, 2022

See also

• Release Management Team