PostgreSQL 16 Open Items

Open Issues

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

NOTE: If known, please list the Owner of the open item.

• Switch to ICU for 17?
  • Owner: Jeff Davis
  • I suggest waiting until next week to commit it and then see what happens
  • CF Entry
• psql: Add role's membership options to the \du+ command
  • CF Entry
  • NOTE: This is not a committed feature for v16
• The rules for choosing default ICU locale seem pretty unfriendly
  • Owner: Jeff Davis
• could not extend file "base/5/3501" with FileFallocate(): Interrupted system call

• Cleaning up nbtree after logical decoding on standby work
  • Owner: Peter Geoghegan, Andres Freund
  • Original commit: 61b313e4
• Inconsistent nulling bitmap in nestloop parameters
  • Owner: Tom Lane

Decisions to Recheck Mid-Beta

• Reconsider a utility_query_id GUC to control if query jumbling of utilities can go through the past string-only mode and the new mode?
  • Potential owner: Michael Paquier

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 Another report Another report

• 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.

• Requiring recovery.signal or standby.signal when recovering with a backup_label
  • This is a rather old behavior that affects all stable branches, still not something that should be backpatched as-is.

• pg_visibility's pg_check_visible() yields false positive when working in parallel with autovacuum
  • Thread with patch CF Entry

• We are not compatible with newly-released LLVM 16
  • Patch
  • Owner: Thomas Munro (volunteer LLVM API change chaser)

• DROP DATABASE is interruptible
  • Additional discussion: "PANIC: could not open critical system index 2662" - twice

Fixed issues

• 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.
  • Fixed at: 8fcb32db98eda1ad2a0c0b40b1cbb5d9a7aa68f0 and ffd1b6bb6f8a2ffc929699772610c6925364dbb3 for HEAD.

• Fix fseek() detection of unseekable files on WIN32
  • Fixed at: a923e21631a29dc8b8781d7d02b5003d0df64ca3 and 765f5df726918bcdcfd16bcc5418e48663d1dd59, down to 14.

• Can't disable autovacuum cost delay through storage parameter
  • Fixed at: bfac8f8bc4a44c67c9f35b5266676278e4ba1217, down to 11.

• EPQ misbehaves for inherited/partitioned tables
  • Fixed at: 70b42f279, down to 14.

Non-bugs

• WindowAgg startup costs don't take into account partition bound. Can lead to incorrect use of cheap startup plans
  • Patch to fix and discussion

Resolved Issues

resolved before 16beta2

• Assert failure and wrong query results due to incorrectly removing PHV
  • Owner: Tom Lane
  • Fixed at: 9a2dbc614e6e47da3c49daacec106da32eba9467
• Assert failure of the cross-check for nullingrels
  • Owner: Tom Lane
  • Original commit: 2489d76c4
  • CF Entry
  • Fixed at: 991a3df22

resolved before 16beta1

• ERROR: wrong varnullingrels (b 5 7) (expected (b)) for Var 3/3
  • Fixed at: d0f952691
• Possible regression setting GUCs on \connect
  • Owner: Alexander Korotkov
  • Discussion on reverting 096dd80f3
  • Original commit: 096dd80f3
  • Reverted at: b9a7a822723aebb16cbe7e5fb874e5124745b07e

• Planner makes improper clause pushdown decisions due to outer-join-aware-Vars changes
  • Clause accidentally pushed down
  • wrong results due to qual pushdown
  • Original commit: 2489d76c4
  • Fixed at: 9df8f903eb6758be5a19e66cdf77e922e9329c31

• Revert ec386948948, per Re: "variable not found in subplan target list"
  • Reverted at 5472743d9e8

• ERROR: ORDER/GROUP BY expression not found in targetlist
  • Fixed at: da5800d5fa636c6e10c9c98402d872c76aa1c8d0

• various elogs hit by sqlsmith (ExecRTCheckPerms() and many prunable partitions)
  • Fixed at: c7468c73f7b6e842a53c12eaee5578a76a8fa7a6

• pg_dump: zlib compression fails for empty objects (LOs)
  • Fixed at: 00d9dcf5bebbb355152a60f0e2120cdf7f9e7ddd

• pg_dump: lz4 compression uses no persistent state and writes a block header for every row
  • Fixed at: 0070b66fef21e909adb283f7faa7b1978836ad75

• Assertion failure with parallel full hash join
  • Fixed at: b37d051b0e59e4324e346655a27509507813db79

• De-revert "Add support for Kerberos credential delegation"
  • Owner: Stephen Frost
  • Original commit: 3d4fa227bce4294ce1cc214b4a9d3b7caa3f0454
  • Revert: (3d03b24c350ab060bb223623bdff38835bd7afd0
  • De-Revert: 6633cfb21691840c33816a6dacaca0b504efb895
  • Resolved at: f7431bca8b0138bdbce7025871560d39119565a0

• SSL tests break on non-existing system CA pool
  • Fixed at: 0b5d1fb36adda612bd3d5d032463a6eeb0729237

• VacuumUpdateCosts() logging condition incorrect for some initial values of vacuum_cost_delay
  • Fixed at: a9781ae11ba2fdb44a3a72c9a7ebb727140b25c5

• check_strxfrm_bug()
  • Owner: Thomas Munro
  • Fixed at: 7d3d72b55edd1b7552a9a358991555994efab0e9

• Should we remove vacuum_defer_cleanup_age?
  • Owner: Andres Freund
  • Fixed at: 1118cd37eb61e6a2428f457a8b2026a7bb3f801a

• Add two missing tests in 035_standby_logical_decoding.pl
  • Fixed at: 376dc820531bafcbf105fff74c5b14c23d9950af
  • Fixed at: a6e04b1d20c2e9cece9b64bb5b36ebfdc3a9031b

• Performance regression due to SQLValueFunction removal
  • Fixed at: d8c3106bb60e4f87be595f241e173ba3c2b7aa2c

• pg_stat_io not tracking smgrwriteback() is confusing
  • Owner: Andres Freund
  • Fixed at: 093e5c57d506783a95dd8feddd9a3f2651e1aeba

• Table data compression is broken with pg_dump --compress lz4
  • Owner: Tomas Vondra
  • Fixed at: 1a05c1d252993b0a59c58a6daf91a2df9333044f

• Valgrind unhappy with LZ4F code in pg_dump
  • Owner: Tomas Vondra
  • Fixed at: 3c18d90f8907e53c3021fca13ad046133c480e4d

• walsender performance regression due to logical decoding on standby changes
  • Owner: Andres Freund
  • Original commit: e101dfac
  • Fixed at: bc971f4025c378ce500d86597c34b0ef996d4d8c

Won't Fix

• Is it OK that WL_SOCKET_ACCEPT is less fair on Windows than on Unix (and than the coding before 16) when there are multiple server sockets configured?
  • WL_SOCKET_ACCEPT fairness on Windows has a (blind) patch to fix that, but would need a Windows hacker to test
  • Owner: Thomas Munro
  • Original commit: 7389aad6
  • Issue reclassified as a non-critical improvement to be considered for 17

Important Dates

Current schedule:

• Beta 2: TBD
• Beta 1: May 25, 2023
• Feature Freeze: April 8, 2023 0:00 AoE (Last Day to Commit Features)

See also

• Release Management Team
• PostgreSQL 15 Open Items