| From: | "Andrey M(dot) Borodin" <x4mmm(at)yandex-team(dot)ru> |
|---|---|
| To: | Heikki Linnakangas <hlinnaka(at)iki(dot)fi> |
| Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: pg_rewind is not crash safe |
| Date: | 2020-08-06 04:55:51 |
| Message-ID: | 8604C380-E070-4853-A6FA-D20AFA0B918B@yandex-team.ru |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> 5 авг. 2020 г., в 23:13, Heikki Linnakangas <hlinnaka(at)iki(dot)fi> написал(а):
>
> A colleague of mine brought to my attention that pg_rewind is not crash safe. If it is interrupted for any reason, it leaves behind a data directory with a mix of data from the source and target images. If you're "lucky", the server will start up, but it can be in an inconsistent state.
FWIW we routinely encounter cases when after unsuccessful pg_rewind databases refuses to start with "contrecord requested" message.
I did not investigate this in detail yet, but I think it is a result of wrong redo recptr written to control file (due to interruption or insufficient present WAL segments).
Best regards, Andrey Borodin.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Noah Misch | 2020-08-06 05:00:02 | Re: public schema default ACL |
| Previous Message | Andrey M. Borodin | 2020-08-06 04:50:40 | Re: Amcheck: do rightlink verification with lock coupling |