From: | Ahmed Bessifi <ahmed(dot)bessifi(at)gmail(dot)com> |
---|---|
To: | pgsql-admin(at)postgresql(dot)org |
Subject: | incorrect pgbench results when postgres fails |
Date: | 2014-01-16 08:53:39 |
Message-ID: | CAJJHcgQCrx9=RU7agXT4-Cw66j0DWzFgBfM9pXiKBdRTGdxG3Q@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Hi,
I recently used pgbench to start some perf tests with pgpool and two
postgres servers set up in master/slave mode.
In fact, during a benchmark (with pgbench) I stopped the slave (test
failover) node and here the test finish quickly (not as expected) and I get
a faulty 'tps' results :
(Bad result)
tps = 202.508908 (including connections establishing)
tps = 210.639230 (excluding connections establishing)
Instead of :
(Good result)
tps = 55.020587 (including connections establishing)
tps = 55.044625 (excluding connections establishing)
I think
this is
a bug
, and I want just inform you about it and see if it could be reported.
Best regards,
Ahmed
--
*"*
*We build software for people who build software.."--Ahmed Bessifi*
--
DevOps Engineer in IMDEV
Engineer degree in computer science from FST since 2013
Degree in system and network administration from IsetCom since 2010
Member in CLLFST club.
Member in open vSwitch community
From | Date | Subject | |
---|---|---|---|
Next Message | Bernhard Schrader | 2014-01-16 13:46:13 | clog problems after dump/restore |
Previous Message | Little, Doug | 2014-01-15 18:49:14 | hiding encrypt/decrypt password |