From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Himanshu Baweja <himanshubaweja(at)yahoo(dot)com> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: how to calculate checkpoint_segments |
Date: | 2005-05-11 16:21:16 |
Message-ID: | 5404.1115828476@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Himanshu Baweja <himanshubaweja(at)yahoo(dot)com> writes:
> i am trying to optimise postgres 8 running on a system.... is there any way to know how wht should be the value of checkpoint_segments...
Enough so you usually aren't checkpointing more often than is specified
by checkpoint_timeout. If you do not know what your system's normal
consumption of WAL is, try setting checkpoint_warning to the same value
as checkpoint_timeout and then keep an eye on the postmaster log to see
how often it complains. You really really *don't* want the thing
checkpointing more often than once every five or ten or so minutes.
> increasing checkpoint_segments degrade the performace while checkpointing as it will have a whole lot of dirty buffers to write.....
This is a fundamental misconception --- the bgwriter exists to prevent that.
See the pgsql-performance archives for more information.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Christopher Murtagh | 2005-05-11 16:40:45 | Re: Trigger that spawns forked process |
Previous Message | Alban Hertroys | 2005-05-11 16:16:19 | Re: Disabling Triggers |