Horton

Re: MultiXact\SLRU buffers configuration

Robert Haas <robertmhaas@gmail.com>

View thread View on postgr.es
From: Robert Haas <robertmhaas@gmail.com>
To: Thom Brown <thom@linux.com>
Cc: Michael Paquier <michael@paquier.xyz>, Alvaro Herrera <alvherre@alvh.no-ip.org>, "Andrey M. Borodin" <x4mmm@yandex-team.ru>, vignesh C <vignesh21@gmail.com>, Andrew Borodin <amborodin86@gmail.com>, Yura Sokolov <y.sokolov@postgrespro.ru>, Andres Freund <andres@anarazel.de>, Thomas Munro <thomas.munro@gmail.com>, Gilles Darold <gilles@darold.net>, Alexander Korotkov <aekorotkov@gmail.com>, Daniel Gustafsson <daniel@yesql.se>, Kyotaro Horiguchi <horikyota.ntt@gmail.com>, pgsql-hackers mailing list <pgsql-hackers@postgresql.org>
Date: 2024-11-05T15:58:37Z
Lists: pgsql-hackers 
On Tue, Oct 29, 2024 at 1:45 PM Thom Brown <thom@linux.com> wrote:
> Taking a look at what's happening under the hood, it seems to be
> getting stuck here:
>
>                 if (nextMXOffset == 0)
>                 {
>                         /* Corner case 2: next multixact is still
> being filled in */
>                         LWLockRelease(MultiXactOffsetSLRULock);
>                         CHECK_FOR_INTERRUPTS();
>                         pg_usleep(1000L);
>                         goto retry;
>                 }
>
> It clearly checks for interrupts, but when I saw this issue happen, it
> wasn't interruptible.

I don't understand the underlying issue here; however, if a process
holds an lwlock, it's not interruptible, even if it checks for
interrupts. So it could be that at this point in the code we're
holding some other LWLock, such as a buffer content lock, and that's
why this code fails to achieve its objective.

-- 
Robert Haas
EDB: http://www.enterprisedb.com

Commits

  1. Rework new SLRU test with injection points

  2. injection_point: Add injection_points.stats

  3. injection_points: Add initialization of shmem state when loading module

  4. Add injection-point test for new multixact CV usage

  5. Use conditional variable to wait for next MultiXact offset

  6. Improve performance of subsystems on top of SLRU

  7. Adjust VACUUM hastup LP_REDIRECT comments.

  8. Test replay of regression tests, attempt II.