Re: wal_buffers, redux
Robert Haas <robertmhaas@gmail.com>
From: Robert Haas <robertmhaas@gmail.com>
To: Andres Freund <andres@anarazel.de>
Cc: pgsql-hackers@postgresql.org, Jeff Janes <jeff.janes@gmail.com>
Date: 2012-03-13T22:23:33Z
Lists: pgsql-hackers
Attachments
- tps-16MB-10s.png (image/png)
On Tue, Mar 13, 2012 at 4:55 AM, Andres Freund <andres@anarazel.de> wrote: > On Tuesday, March 13, 2012 03:26:34 AM Robert Haas wrote: >> Meanwhile, here are some TPS graphs at 16MB and 32MB on the IBM POWER7 >> machine. 32 clients, 1800 seconds, scale factor 300, synchronous >> commit off. > That graph makes me cringe because its pretty representative of what I have > seen in practise. Any chance we can get a ~10s average in there? I find it > pretty hard to read more than "spikey". I tried that but don't find it an improvement; however I'm attaching an example for your consideration. > Whats the filesystem + options on this? ext4, default options. Fedora 16, 3.2.6-3.fc16.ppc64. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company