From: Tom Lane <tgl@sss.pgh.pa.us>
To: exclusion@gmail.com
cc: Peter Eisentraut <peter.eisentraut@enterprisedb.com>,
        pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #17416: Server crashes due to python3 stack overflow on
 executing multiple plpy.rollback() calls
In-reply-to: <17416-ed8fe5d7213d6c25@postgresql.org>
References: <17416-ed8fe5d7213d6c25@postgresql.org>
Comments: In-reply-to PG Bug reporting form <noreply@postgresql.org>
	message dated "Wed, 23 Feb 2022 11:00:01 +0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1048490.1645633650.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Wed, 23 Feb 2022 11:27:30 -0500
Message-ID: <1048491.1645633650@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1048491.1645633650%40sss.pgh.pa.us>
Precedence: bulk

PG Bug reporting form <noreply@postgresql.org> writes:
> When executing the following script:
> psql -c "
> CREATE EXTENSION plpython3u;

> CREATE FUNCTION plpy_xact_test() RETURNS int
> LANGUAGE plpython3u
> AS '
> plpy.rollback()
> return 1
> '"

> for i in `seq 500`; do echo "SELECT plpy_xact_test();"; done | psql

> I get the server crash with the following stack trace:

Interesting.  For me, 500 repetitions won't crash it, but 1000 will.
Apparently, each escape out of plpy.rollback leaves an open frame
on Python's internal control stack, and eventually it barfs.
So this is a different way to reach the problems already under
investigation at [1].  You don't really need a bleeding-edge
Python at all.

			regards, tom lane

[1] https://www.postgresql.org/message-id/flat/3375ffd8-d71c-2565-e348-a59=
7d6e739e3@enterprisedb.com