Something is wrong with wal_compression

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: pgsql-hackers(at)lists(dot)postgresql(dot)org
Cc: Michael Paquier <michael(at)paquier(dot)xyz>
Subject: Something is wrong with wal_compression
Date: 2023-01-26 19:43:29
Message-ID: 3163112.1674762209@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

The symptom being exhibited by Michael's new BF animal tanager
is perfectly reproducible elsewhere.

$ cat /home/postgres/tmp/temp_config
#default_toast_compression = lz4
wal_compression = lz4
$ export TEMP_CONFIG=/home/postgres/tmp/temp_config
$ cd ~/pgsql/src/test/recovery
$ make check PROVE_TESTS=t/011_crash_recovery.pl
...
+++ tap check in src/test/recovery +++
t/011_crash_recovery.pl .. 1/?
# Failed test 'new xid after restart is greater'
# at t/011_crash_recovery.pl line 53.
# '729'
# >
# '729'

# Failed test 'xid is aborted after crash'
# at t/011_crash_recovery.pl line 57.
# got: 'committed'
# expected: 'aborted'
# Looks like you failed 2 tests of 3.

Maybe this is somehow the test script's fault, but I don't see how.

It fails the same way with 'wal_compression = pglz', so I think it's
generic to that whole feature rather than specific to LZ4.

regards, tom lane

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Nathan Bossart 2023-01-26 19:48:12 Re: suppressing useless wakeups in logical/worker.c
Previous Message Drouvot, Bertrand 2023-01-26 19:43:25 Re: Helper functions for wait_for_catchup() in Cluster.pm