Horton

Re: pgbench logging broken by time logic changes

Fabien COELHO <coelho@cri.ensmp.fr>

View thread View on postgr.es
From: Fabien COELHO <coelho@cri.ensmp.fr>
To: Michael Paquier <michael@paquier.xyz>
Cc: Alvaro Herrera <alvherre@alvh.no-ip.org>, Andrew Dunstan <andrew@dunslane.net>, Thomas Munro <thomas.munro@gmail.com>, Gregory Smith <gregsmithpgsql@gmail.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>, david.christensen@crunchydata.com
Date: 2021-06-24T06:46:03Z
Lists: pgsql-hackers

Attachments

Bonjour Michaël,

> Using grep() with "$re" results in all the fields matching.  Using on
> the contrary "/$re/" in grep(), like list_files(), would only match
> the first one, which is correct.

Ok, good catch. Perl is kind of a strange language.

> With this issue fixed, I have bumped into what looks like a different 
> bug in the tests.  001_pgbench_log_2 uses pgbench with 2 clients, but> 
> expects only patterns in the logs where the first column value uses only 
> 0.  With two clients, those first values can be either 0 or 1 due to the 
> client ID set.

Indeed. The tests passes because the number of expected lines is quite

> It seems to me that we had better fix this issue and back-patch where
> this has been introduced so as we have exact match checks with the log
> formarts, no?  Please see the attached.

Ok, however the regex should be "^[01] ...".

Attached v11 with your fixes + the above regex fix.

-- 
Fabien.

Commits

  1. Fix pgbench timestamp bugs.

  2. Fix pattern matching logic for logs in TAP tests of pgbench

  3. pgbench: Improve time logic.

  4. pgbench: Synchronize client threads.

  5. pgbench: refactor handling of stats tracking