Content-Type: text/plain;
  charset="iso-8859-1"
From: Richard Huxton <dev@archonet.com>
Organization: Archonet Ltd
To: Wei Weng <wweng@kencast.com>
Subject: Re: concurrent connections is worse than serialization?
Date: Wed, 14 Aug 2002 15:49:57 +0100
User-Agent: KMail/1.4.1
Cc: pgsql-sql@postgresql.org
References: <1029271170.16828.7.camel@Monet>
	<200208141018.38875.dev@archonet.com>
	<1029334827.21470.2.camel@Monet>
In-Reply-To: <1029334827.21470.2.camel@Monet>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Message-Id: <200208141549.57646.dev@archonet.com>

On Wednesday 14 Aug 2002 3:20 pm, Wei Weng wrote:
> On Wed, 2002-08-14 at 05:18, Richard Huxton wrote:
> > On Tuesday 13 Aug 2002 9:39 pm, Wei Weng wrote:

[30 connections is much slower than 1 connection 30 times]

> > What was the limiting factor during the test? Was the CPU maxed, memory,
> > disk I/O?
>
> No, none of the above was maxed. CPU usage that I paid attention to was
> at most a 48%.

Something must be the limiting factor. One of
 - CPU
 - Memory
 - Disk I/O
 - Database (configuration, or design)
 - Application

If it's not CPU, is the system going into swap or are you seeing a lot of d=
isk=20
activity?

> > I assume you've ruled the application end of things out.
>
> What does this mean?

I mean if you don't actually run the queries, then 30 separate processes is=
=20
fine?

If you can provide us with an EXPLAIN of the query and the relevant schema=
=20
definitions, we can rule out database design.

- Richard Huxton