| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Josh Berkus <josh(at)agliodbs(dot)com> |
| Cc: | pgsql-performance(at)postgresql(dot)org, "John Parnefjord" <John(dot)Parnefjord(at)kib(dot)ki(dot)se> |
| Subject: | Re: Tuning |
| Date: | 2007-01-28 23:38:18 |
| Message-ID: | 26422.1170027498@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
Josh Berkus <josh(at)agliodbs(dot)com> writes:
>> -checkpoint_segments - this is crucial as one of the server is
>> transaction heavy
> Well, it only helps you to raise this if you have a dedicated disk resource
> for the xlog. Otherwise having more segments doesn't help you much.
Au contraire, large checkpoint_segments is important for write-intensive
workloads no matter what your disk layout is. If it's too small then
you'll checkpoint too often, and the resulting increase in page-image
writes will hurt. A lot.
My advice is to set checkpoint_warning to the same value as
checkpoint_timeout (typically 5 minutes or so) and then keep an eye on
the postmaster log for awhile. If you see more than a few "checkpoints
are occuring too frequently" messages, you want to raise
checkpoint_segments.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Campbell, Lance | 2007-01-29 03:10:04 | work-mem how do I identify the proper size |
| Previous Message | Josh Berkus | 2007-01-28 23:24:24 | Re: Tuning |