| From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
|---|---|
| To: | Floris Van Nee <florisvannee(at)optiver(dot)com> |
| Cc: | PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: visibility map corruption |
| Date: | 2021-07-04 21:52:07 |
| Message-ID: | CAH2-Wzk2T16WtN56MJEkOcW2UN9Hzn7JNMSPERgRL_xZv5Xocg@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Sun, Jul 4, 2021 at 2:26 PM Floris Van Nee <florisvannee(at)optiver(dot)com> wrote:
> > Have you ever used pg_upgrade on this database?
> >
>
> Yes. The last time (from v11 to v12) was in October 2020. The transaction id in the tuples (the one PG is trying to check in the tx log) dated from February 2021. I do believe (but am not 100% certain) that the affected table already existed at the time of the last pg_upgrade though.
I wonder if it's related to this issue:
https://www.postgresql.org/message-id/20210423234256.hwopuftipdmp3okf@alap3.anarazel.de
Have you increased autovacuum_freeze_max_age from its default? This
already sounds like the kind of database where that would make sense.
--
Peter Geoghegan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Justin Pryzby | 2021-07-04 22:22:59 | Re: "debug_invalidate_system_caches_always" is too long |
| Previous Message | Floris Van Nee | 2021-07-04 21:26:00 | RE: visibility map corruption |