Horton

Re: BUG #16678: The ecpg connect/test5 test sometimes fails on Windows

Kyotaro Horiguchi <horikyota.ntt@gmail.com>

View thread View on postgr.es
From: Kyotaro Horiguchi <horikyota.ntt@gmail.com>
To: exclusion@gmail.com
Cc: tgl@sss.pgh.pa.us, pgsql-bugs@lists.postgresql.org
Date: 2020-10-26T01:12:18Z
Lists: pgsql-bugs 
At Sun, 25 Oct 2020 08:00:00 +0300, Alexander Lakhin <exclusion@gmail.com> wrote in 
> And regarding the src/test/recovery test failure with a 10 ms delay
> added... I've increased a log level and get probably the same behaviour
> as described in
> https://www.postgresql.org/message-id/flat/OFCA523F90.7499E22F-ON4325828E.005573C4-4325828E.0055B412%40iba.by
> The primary node can't finish smart shutdown and logs of the primary and
> the standby are filling with the following messages:
...
>  
> Though the patch proposed by Kyotaro Horiguchi doesn't help. It this
> looks like an issue that should be fixed, I can investigate it further.

Yeah, that patch just rejects to start from a wrong backup. If a
specific legit operation makes such a backup, it should be fixed.

regareds.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center

Commits

  1. On Windows, close the client socket explicitly during backend shutdown.