From: Antonin Houska <ah@cybertec.at>
To: pgsql-hackers@lists.postgresql.org
Subject: Re: Race conditions in logical decoding
In-reply-to: <85833.1768840165@localhost>
References: <85833.1768840165@localhost>
Comments: In-reply-to Antonin Houska <ah@cybertec.at>
   message dated "Mon, 19 Jan 2026 17:29:25 +0100."
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="=-=-="
Date: Tue, 20 Jan 2026 09:30:33 +0100
Message-ID: <62335.1768897833@localhost>
Archived-At: <https://www.postgresql.org/message-id/62335.1768897833%40localhost>
Precedence: bulk

--=-=-=
Content-Type: text/plain

Antonin Houska <ah@cybertec.at> wrote:

> I'm not sure yet how to fix the problem. I tried to call XactLockTableWait()
> from SnapBuildAddCommittedTxn() (like it happens in SnapBuildWaitSnapshot()),
> but it made at least one regression test (subscription/t/010_truncate.pl)
> stuck - probably a deadlock. I can spend more time on it, but maybe someone
> can come up with a good idea sooner than me.

Attached here is what I consider a possible fix - simply wait for the CLOG
update before building a new snapshot.

Unfortunately I have no idea right now how to test it using the isolation
tester. With the fix, the additional waiting makes the current test
block. (And if a step is added that unblock the session, it will not reliably
catch failure to wait.)

-- 
Antonin Houska
Web: https://www.cybertec-postgresql.com


--=-=-=
Content-Type: text/x-diff
Content-Disposition: attachment;
 filename=0001-Fix-race-conditions-during-the-setup-of-logical-deco.patch