From: Tom Lane <tgl@sss.pgh.pa.us>
To: higherone@gmail.com
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #17602: Query backend process killed because it uses up all
 memory.
In-reply-to: <17602-b072900fa964cabf@postgresql.org>
References: <17602-b072900fa964cabf@postgresql.org>
Comments: In-reply-to PG Bug reporting form <noreply@postgresql.org>
	message dated "Tue, 30 Aug 2022 16:11:40 -0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <3727392.1661886091.1@sss.pgh.pa.us>
Date: Tue, 30 Aug 2022 15:01:31 -0400
Message-ID: <3727393.1661886091@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/3727393.1661886091%40sss.pgh.pa.us>
Precedence: bulk

PG Bug reporting form <noreply@postgresql.org> writes:
> I'm writing to seek for generic help on trouble shooting bugs related to
> query backend process killed because of memory exhaust on this single query
> process.

Try starting the server under a restrictive ulimit, so that it will
get ENOMEM failure before it bloats to the point of receiving an
OOM kill.  Aside from being more friendly to other server processes,
this should result in getting a memory consumption map dumped to
postmaster stderr when the failure is hit.  Seeing which context
is getting bloated would be pretty useful information.

On Linux I believe "ulimit -v" is the most appropriate setting to
change.

			regards, tom lane