Re: Extend documentation for pg_stat_replication.backend_xmin

Fujii Masao <masao.fujii@gmail.com>

From: Fujii Masao <masao.fujii@gmail.com>
To: Renzo Dani <arons7@gmail.com>
Cc: pgsql-hackers@postgresql.org
Date: 2025-10-16T03:31:57Z
Lists: pgsql-hackers
On Thu, Oct 16, 2025 at 12:47 AM Renzo Dani <arons7@gmail.com> wrote:
>
> Hi,
> recently I had some issue understanding what was going on with a replica and the reported xmin.
> I have than created a patch to better specify the value expected into the pg_stat_replication.backend_xmin field.

+1

+       This field will be null if a replication slot is used; in that
case, the xmin is available through <link
linkend="view-pg-replication-slots">pg_replication_slots</link>.

I think it would be better to also mention that this field is null
when hot_standby_feedback is disabled.

Also, <literal> tags should be used for "xmin", and a newline should
be added to avoid an overly long line.

Based on the above comments, how about revising it as follows?

+       This field will be null if a replication slot is used (in this case,
+       the standby's <literal>xmin</literal> is shown in
+       <link linkend="view-pg-replication-slots">pg_replication_slots</link>)
+       or if <varname>hot_standby_feedback</varname> is disabled.

Regards,

-- 
Fujii Masao



Commits

  1. doc: Clarify when backend_xmin in pg_stat_replication can be NULL.