Re: pg_reset_stats + cache I/O %

From: "Jim C(dot) Nasby" <jnasby(at)pervasive(dot)com>
To: "mcelroy, tim" <tim(dot)mcelroy(at)bostonstock(dot)com>
Cc: "'pgsql-performance(at)postgresql(dot)org'" <pgsql-performance(at)postgresql(dot)org>
Subject: Re: pg_reset_stats + cache I/O %
Date: 2006-03-09 03:23:37
Message-ID: 20060309032337.GW45250@pervasive.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

On Wed, Mar 08, 2006 at 01:35:35PM -0500, mcelroy, tim wrote:
> I actually need this info as I was tasked by management to provide it. Not
> sure if they understand that or not, I do but management does like to see
> how well the system and its components are performing. Also, I would
> utilize these results to test any cache tuning changes I may make.

What I feared. While monitoring cache hit % over time isn't a bad idea,
it's less than half the picture, which makes fertile ground for
optimizing for some mythical target instead of actual system
performance. If the "conclusion" from these numbers is that
shared_buffers needs to get set larger than min(50000, 10% of memory)
I'd very seriously re-consider how performance tuning is being done.

But hopefully I'm just being paranoid and you guys are just doing a
great job of monitoring things and keeping on the ball. :)
--
Jim C. Nasby, Sr. Engineering Consultant jnasby(at)pervasive(dot)com
Pervasive Software http://pervasive.com work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf cell: 512-569-9461

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Javier Somoza 2006-03-09 10:24:36 pgCluster and postgres 8.1
Previous Message Gorshkov 2006-03-09 00:32:37 embedded postgres and threading