| From: | Andrey Borodin <x4mmm(at)yandex-team(dot)ru> |
|---|---|
| To: | Melanie Plageman <melanieplageman(at)gmail(dot)com> |
| Cc: | Andres Freund <andres(at)anarazel(dot)de>, Matthias van de Meent <boekewurm+postgres(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Robert Haas <robertmhaas(at)gmail(dot)com>, Heikki Linnakangas <hlinnaka(at)iki(dot)fi> |
| Subject: | Re: Why clearing the VM doesn't require registering vm buffer in wal record |
| Date: | 2026-10-06 06:38:02 |
| Message-ID: | 7A3B2E5A-6E83-46BE-8705-EBCB7CFD5B58@yandex-team.ru |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi Melanie,
On 24 Sep 2026, Melanie Plageman wrote:
> v2-0001 is 19-only and fixes the divergent VM bits on the standby
The UPDATE fallback from v1 is missing in both versions of 0001.
heap_xlog_update() handles VM_OLD/VM_NEW itself, without calling
heap_xlog_vm_clear(), so an unregistered VM page is still skipped.
I reproduced this with UPDATE in your CREATE DATABASE/failover scenario
on master. Could we add the fallback there too?
Separately, Shihao's v3 [0] adds PageInit() fix on top of your 0002.
Thank you!
Best regards, Andrey Borodin.
| From | Date | Subject | |
|---|---|---|---|
| Previous Message | Hayato Kuroda (Fujitsu) | 2026-10-06 06:37:32 | RE: Parallel Apply |