From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
cc: pgsql-hackers@postgresql.org, Thomas Munro <thomas.munro@gmail.com>
Subject: Re: failures in t/031_recovery_conflict.pl on CI
In-reply-to: <20220412184913.3wrt4olkdiqvicsj@alap3.anarazel.de>
References: <20220409045515.35ypjzddp25v72ou@alap3.anarazel.de>
 <20220409050501.frejv2i46m54keem@alap3.anarazel.de>
 <20220409220054.fqn5arvbeesmxdg5@alap3.anarazel.de>
 <20220409231002.mddrr7mtca7p55v5@alap3.anarazel.de>
 <20220409233107.v7okovwr46xsizfk@alap3.anarazel.de>
 <5111.1649547266@sss.pgh.pa.us>
 <20220412184913.3wrt4olkdiqvicsj@alap3.anarazel.de>
Comments: In-reply-to Andres Freund <andres@anarazel.de>
	message dated "Tue, 12 Apr 2022 11:49:13 -0700"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <907759.1649790322.1@sss.pgh.pa.us>
Date: Tue, 12 Apr 2022 15:05:22 -0400
Message-ID: <907760.1649790322@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/907760.1649790322%40sss.pgh.pa.us>
Precedence: bulk

Andres Freund <andres@anarazel.de> writes:
> On 2022-04-09 19:34:26 -0400, Tom Lane wrote:
>> +1.  This is probably more feasible given the latch infrastructure
>> than it was when that code was first written.

> What do you think about just reordering the disable_all_timeouts() to be
> before the got_standby_deadlock_timeout check in the back branches? I think
> that should close at least the most obvious hole.  And fix it properly in
> HEAD?

I don't have much faith in that, and I don't see why we can't fix it
properly.  Don't we just need to have the signal handler set MyLatch,
and then do the unsafe stuff back in the "if (got_standby_deadlock_timeout)"
stanza in mainline?

			regards, tom lane