Re: recovery_target_action=pause with confusing hint

From: Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>
To: "movead(dot)li(at)highgo(dot)ca" <movead(dot)li(at)highgo(dot)ca>, Sergei Kornilov <sk(at)zsrv(dot)org>, pgsql-hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: recovery_target_action=pause with confusing hint
Date: 2020-04-08 09:25:13
Message-ID: 7452777f-d7b1-f336-4882-b7802d79b9ab@oss.nttdata.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2020/04/08 16:41, movead(dot)li(at)highgo(dot)ca wrote:
>
> >> we should notice it in document I think.
>>There is the following explation about the relationship the recovery
>>pause and the promotion, in the document. You may want to add more
>>descriptions into the docs?
> >------------------------------
>>If a promotion is triggered while recovery is paused, the paused
>>state ends and a promotion continues.
>>------------------------------
>
> For example we can add this words:
> First-time pg_wal_replay_pause() called during recovery which triggered
> as promotion, pg_wal_replay_pause() show success but it did not really
> pause the recovery.

I think this is not true. In your case, pg_wal_replay_pause() succeeded
because the startup process had not detected the promotion request yet
at that moment.

To cover your case, what about adding the following description?

-----------------
There can be delay between a promotion request by users and the trigger of
a promotion in the server. Note that pg_wal_replay_pause() succeeds
during that delay, i.e., until a promotion is actually triggered.
-----------------

Regards,

--
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2020-04-08 09:26:30 Re: adding partitioned tables to publications
Previous Message Pavel Borisov 2020-04-08 09:13:45 [PATCH] RUM Postgres 13 patch