Message-ID: <5419F1CC.8060309@gmx.net>
Date: Wed, 17 Sep 2014 16:40:44 -0400
From: Peter Eisentraut <peter_e@gmx.net>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9;
 rv:31.0) Gecko/20100101 Thunderbird/31.1.1
MIME-Version: 1.0
To: "Huang, Suya" <Suya.Huang@au.experian.com>,
 "pgsql-admin@postgresql.org" <pgsql-admin@postgresql.org>
Subject: Re: How to interpret view pg_stat_bgwriter
References: 
 <D83E55F5F4D99B4A9B4C4E259E6227CD01EBFA82@AUX1EXC02.apac.experian.local>
In-Reply-To: 
 <D83E55F5F4D99B4A9B4C4E259E6227CD01EBFA82@AUX1EXC02.apac.experian.local>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

On 9/16/14 2:42 AM, Huang, Suya wrote:
> # checkpoints_req is much bigger than checkpoints_timed, suggest that I
> may increase checkpoint_segments in our system

That's not strictly necessary, but personally I think both of those
settings are too low for you.

> #maxwritten_clean is high, suggests increase bgwriter_lru_maxpages

Yes.

> # buffers_backend is much smaller than buffers_alloc, suggests
> increasing bgwriter_lru_maxpages, bgwriter_lru_multiplier, and
> decreasing bgwriter_delay.

You should rather compare buffers_backend against buffers_checkpoint.


Also consider graphing these quantities over time so you can see how
different workloads affects the results and what the effects of your
tuning are.

Also turn on log_checkpoints and look at the sync times.  In my
experience that dominates all the bgwriter tuning.


-- 
Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin