Re: Checkpoint spikes - Mailing list pgsql-performance

From Kevin Grittner
Subject Re: Checkpoint spikes
Date
Msg-id 4B1E16C0020000250002D1D9@gw.wicourts.gov
Whole thread Raw
In response to Re: Checkpoint spikes  (Richard Neill <rn214@cam.ac.uk>)
List pgsql-performance
Richard Neill <rn214@cam.ac.uk> wrote:

> So far, I've set checkpoint_segments to 128, timeout to 10min, and
> completion_target to 0.8. This helps, but not as much as I'd
> hoped.
>
> But I haven't touched any of the other WAL or BG Writer settings.
>
> Where should I look next?

On our web servers, where we had similar issues, we seem to be doing
OK using:

bgwriter_lru_maxpages = 1000
bgwriter_lru_multiplier = 4.0

The other thing which can help this problem is keeping
shared_buffers smaller than most people recommend.  We use 512MB on
our larger web server and 256MB on other servers.  (Be sure to test
with your actual load, as this might or might not degrade overall
performance.)

-Kevin

pgsql-performance by date:

Previous
From: Robert Haas
Date:
Subject: Re: performance penalty between Postgresql 8.3.8 and 8.4.1
Next
From: Robert Haas
Date:
Subject: Re: performance penalty between Postgresql 8.3.8 and 8.4.1