Re: [Bug fix]There is the case archive_timeout parameter is ignored after recovery works.

Hello.

At Mon, 29 Jun 2020 04:35:11 +0000, "higuchi(dot)daisuke(at)fujitsu(dot)com" <higuchi(dot)daisuke(at)fujitsu(dot)com> wrote in
> Hi,
>
> I found the bug about archive_timeout parameter.
> There is the case archive_timeout parameter is ignored after recovery works.
...
> [Problem]
> When the value of archive_timeout is smaller than that of checkpoint_timeout and recovery works, archive_timeout is ignored in the first WAL archiving.
> Once WAL is archived, the archive_timeout seems to be valid after that.
...
> Currently, cur_timeout is set according to only checkpoint_timeout when it is during recovery.
> Even during recovery, the cur_timeout should be calculated including archive_timeout as well as checkpoint_timeout, I think.
> I attached the patch to solve this problem.

Unfortunately the diff command in your test script doesn't show me
anything, but I can understand what you are thinking is a problem,
maybe. But the patch doesn't seem the fix for the issue.

Archiving works irrelevantly from that parameter. Completed WAL
segments are immediately marked as ".ready" and archiver does its task
immediately independently from checkpointer. The parameter, as
described in documentation, forces the server to switch to a new WAL
segment file periodically so that it can be archived, that is, it
works only on primary. On the other hand on standby, a WAL segment is
not marked as ".ready" until any data for the *next* segment comes. So
the patch is not the fix for the issue.

If primary switched segment and archived it but standby didn't archive
the same immediately, you could force that by writing something on the
master.

Anyway, the attached patch would resolve your problem.

regards.

--
Kyotaro Horiguchi
NTT Open Source Software Center