Horton

Re: pgbench: using prepared BEGIN statement in a pipeline could cause an error

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Yugo NAGATA <nagata@sraoss.co.jp>, Tom Lane <tgl@sss.pgh.pa.us>, Fabien COELHO <coelho@cri.ensmp.fr>, pgsql-hackers@postgresql.org
Date: 2023-02-14T00:20:04Z
Lists: pgsql-hackers 
Hi,

On 2023-02-08 13:10:40 +0100, Alvaro Herrera wrote:
> On 2023-Feb-08, Alvaro Herrera wrote:
> 
> > I propose instead the following: each command is prepared just before
> > it's executed, as previously, and if we see a \startpipeline, then we
> > prepare all commands starting with the one just after, and until the
> > \endpipeline.
> 
> Here's the patch.

There's something wrong with the patch, it reliably fails with core dumps:
https://cirrus-ci.com/github/postgresql-cfbot/postgresql/commitfest%2F42%2F3260

Example crash:
https://api.cirrus-ci.com/v1/task/4922406553255936/logs/cores.log
https://api.cirrus-ci.com/v1/artifact/task/6611256413519872/crashlog/crashlog-pgbench.EXE_0750_2023-02-13_14-07-06-189.txt

Andres

Commits

  1. Fix pgbench in prepared mode with an empty pipeline

  2. pgbench: Prepare commands in pipelines in advance

  3. Add -M (query mode) option per ITAGAKI Takahiro