Re: Interpreting vmstat - Mailing list pgsql-performance

From Thom Dyson
Subject Re: Interpreting vmstat
Date
Msg-id OF548594A5.D2D24DD0-ON88256E9A.00575B92-88256E9A.0057ED01@sybex.com
Whole thread Raw
In response to Interpreting vmstat  (Doug Y <dylists@ptd.net>)
List pgsql-performance



Well,

Since I haven't seen any other responds, I'll offer a bit of advice and let
others correct me. :)

Your shared buffers may be too big (?).  It is much larger than the guide
on varlena.com recommends.  All I can suggest is trying some experiments
with halving/doubling the numbers to see which way performance goes.  Also,
if you are counting on cache to improve performance, then the db has to be
loaded into cache the first time. So, are subsequent re-queries faster?

Thom Dyson
Director of Information Services
Sybex, Inc.



pgsql-performance-owner@postgresql.org wrote on 05/18/2004 11:12:14 AM:

> Hello,
> (note best viewed in fixed-width font)
>
> I'm still trying to find where my performance bottle neck is...
> I have 4G ram, PG 7.3.4
> shared_buffers = 75000
> effective_cache_size = 75000



pgsql-performance by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: PostgreSQL performance in simple queries
Next
From: Joseph Shraibman
Date:
Subject: Re: PostgreSQL performance in simple queries