Re: Transaction timeout

From: Alexander Korotkov <aekorotkov(at)gmail(dot)com>
To: "Andrey M(dot) Borodin" <x4mmm(at)yandex-team(dot)ru>
Cc: Japin Li <japinli(at)hotmail(dot)com>, Peter Smith <smithpb2250(at)gmail(dot)com>, Junwang Zhao <zhjwpku(at)gmail(dot)com>, 邱宇航 <iamqyh(at)gmail(dot)com>, Fujii Masao <masao(dot)fujii(at)oss(dot)nttdata(dot)com>, Andrew Borodin <amborodin86(at)gmail(dot)com>, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>, Nikolay Samokhvalov <samokhvalov(at)gmail(dot)com>, pgsql-hackers mailing list <pgsql-hackers(at)postgresql(dot)org>, pgsql-hackers mailing list <pgsql-hackers(at)lists(dot)postgresql(dot)org>, Andres Freund <andres(at)anarazel(dot)de>
Subject: Re: Transaction timeout
Date: 2024-03-11 11:18:01
Message-ID: CAPpHfduKa6rh48LNVG3_X5_7BiAO3+RKtRj-XPWEN2StVkAHLA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Mar 11, 2024 at 12:53 PM Andrey M. Borodin <x4mmm(at)yandex-team(dot)ru> wrote:
> > On 7 Mar 2024, at 00:55, Alexander Korotkov <aekorotkov(at)gmail(dot)com> wrote:
> >
> > On Wed, Mar 6, 2024 at 10:22 AM Andrey M. Borodin <x4mmm(at)yandex-team(dot)ru> wrote:
> >>> On 25 Feb 2024, at 21:50, Alexander Korotkov <aekorotkov(at)gmail(dot)com> wrote:
> >>>
> >>> Thank you for the patches. I've pushed the 0001 patch to avoid
> >>> further failures on buildfarm. Let 0004 wait till injections points
> >>> by Mechael are committed.
> >>
> >> Thanks!
> >>
> >> All prerequisites are committed. I propose something in a line with this patch.
> >
> > Thank you. I took a look at the patch. Should we also check the
> > relevant message after the timeout is fired? We could check it in
> > psql stderr or log for that.
>
> PFA version which checks log output.
> But I could not come up with a proper use of BackgroundPsql->query_until() to check outputs. And there are multiple possible errors.
>
> We can copy test from src/bin/psql/t/001_basic.pl:
>
> # test behavior and output on server crash
> my ($ret, $out, $err) = $node->psql('postgres',
> "SELECT 'before' AS running;\n"
> . "SELECT pg_terminate_backend(pg_backend_pid());\n"
> . "SELECT 'AFTER' AS not_running;\n");
>
> is($ret, 2, 'server crash: psql exit code');
> like($out, qr/before/, 'server crash: output before crash');
> ok($out !~ qr/AFTER/, 'server crash: no output after crash');
> is( $err,
> 'psql:<stdin>:2: FATAL: terminating connection due to administrator command
> psql:<stdin>:2: server closed the connection unexpectedly
> This probably means the server terminated abnormally
> before or while processing the request.
> psql:<stdin>:2: error: connection to server was lost',
> 'server crash: error message’);
>
> But I do not see much value in this.
> What do you think?

I think if checking psql stderr is problematic, checking just logs is
fine. Could we wait for the relevant log messages one by one with
$node->wait_for_log() just like 040_standby_failover_slots_sync.pl do?

------
Regards,
Alexander Korotkov

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Dagfinn Ilmari Mannsåker 2024-03-11 11:19:16 Re: Using the %m printf format more
Previous Message Amit Kapila 2024-03-11 11:06:18 Re: Regardign RecentFlushPtr in WalSndWaitForWal()