From: Tom Lane <tgl@sss.pgh.pa.us>
To: Josef Machytka <josef.machytka@gmail.com>
cc: pgsql-bugs@postgresql.org
Subject: Re: BUG #14471: PostgreSQL 9.6 "NOT IN" in select causes crash with
 "ERROR: unknown error"
In-reply-to: 
 <CAGvVEFt8Gb2ANBb2RnLMJMagyBMDpP7ebYjHoZ5WRG0eB6D5qg@mail.gmail.com>
References: <20161221094443.25614.47807@wrigleys.postgresql.org>
 <6818.1482333970@sss.pgh.pa.us>
 <CAGvVEFs1QxgFLUmc2aasrLYQS5MQhW_v4oV13i15Mph9Cwiugg@mail.gmail.com>
 <18191.1482336664@sss.pgh.pa.us>
 <CAGvVEFt8Gb2ANBb2RnLMJMagyBMDpP7ebYjHoZ5WRG0eB6D5qg@mail.gmail.com>
Comments: In-reply-to Josef Machytka <josef.machytka@gmail.com>
	message dated "Fri, 23 Dec 2016 10:08:42 +0100"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <27431.1482508184.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 23 Dec 2016 10:49:44 -0500
Message-ID: <27432.1482508184@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-bugs-owner@postgresql.org

Josef Machytka <josef.machytka@gmail.com> writes:
> So I checked all logs and it turned out one of databases involved vent in=
to
> recovery mode because some of its connections was killed.
> Probably out of memory killer did it. So it caused chain reaction. And
> billing query was canceled "due to administration command" and it was sho=
wn
> through dblink as "unknown error".

OK.  In that case I'm betting that what dblink saw was a network error on
its next attempt to send a command.  I've not dug into the code to verify
this positively, but I think the behavior will be improved by Joe Conway's
recent patch,
https://git.postgresql.org/gitweb/?p=3Dpostgresql.git;a=3Dcommitdiff;h=3D2f=
802d95b4904dbed3dfdca1b3a607cd085d2e20

			regards, tom lane


--=20
Sent via pgsql-bugs mailing list (pgsql-bugs@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-bugs