Re: minor bug - Mailing list pgsql-hackers

From Laurenz Albe
Subject Re: minor bug
Date
Msg-id b41b61c8ba52efa23d1079cc53c689713f5b1e0c.camel@cybertec.at
Whole thread Raw
Responses Re: minor bug
Re: minor bug
List pgsql-hackers
On Mon, 2023-01-16 at 19:59 +0100, Torsten Förtsch wrote:
> not sure if this is known behavior.
>
> Server version is 14.6 (Debian 14.6-1.pgdg110+1).
>
> In a PITR setup I have these settings:
>
> recovery_target_xid = '852381'
> recovery_target_inclusive = 'false'
>
> In the log file I see this message:
>
> LOG:  recovery stopping before commit of transaction 852381, time 2000-01-01 00:00:00+00
>
> But:
>
> postgres=# select * from pg_last_committed_xact();
>   xid   |           timestamp           | roident
> --------+-------------------------------+---------
>  852380 | 2023-01-16 18:00:35.054495+00 |       0
>
> So, the timestamp displayed in the log message is certainly wrong.

Redirected to -hackers.

If recovery stops at a WAL record that has no timestamp, you get this
bogus recovery stop time.  I think we should show the recovery stop time
only if time was the target, as in the attached patch.

Yours,
Laurenz Albe

Attachment

pgsql-hackers by date:

Previous
From: "shiy.fnst@fujitsu.com"
Date:
Subject: Update comments in multixact.c
Next
From: Antonin Houska
Date:
Subject: Re: Cross-partition UPDATE and foreign table partitions