Re: Recent pg_rewind test failures in buildfarm

Hello Tom,

15.04.2025 05:58, Tom Lane wrote:
> In the last day or so, both skink and mamba have hit this
> in the pg_rewind test suite [1][2]:
>
> #3 0x01f03f7c in ExceptionalCondition (conditionName=conditionName(at)entry=0x2119c4c "pending_since == 0", fileName=fileName(at)entry=0x2119454 "pgstat.c", lineNumber=lineNumber(at)entry=734) at assert.c:66
> ...
>
> That assert appears to be several years old, and the
> 008_min_recovery_point.pl test script that's triggering it hasn't
> changed very recently either, so I'm baffled where to start digging.
> It has the odor of a timing problem, so maybe we just started hitting
> this by chance. Still ... anybody have an idea?
>
> [1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=skink&dt=2025-04-13%2018%3A55%3A03
> [2] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=mamba&dt=2025-04-15%2001%3A00%3A04
>

FWIW, there was also another test (028_row_filter.pl) failed due to that
assert triggered: [3].

I've managed to reproduce this failure locally when running 10 instances
of 008_min_recovery_point.pl in parallel under Valgrind.

Will try to investigate the issue during this week.

[3] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=skink&dt=2025-04-11%2007%3A41%3A36

Best regards,
Alexander Lakhin
Neon (https://neon.tech)