Received: from localhost (unknown [200.46.204.182])
	by postgresql.org (Postfix) with ESMTP id 8285D9F96CE
	for <pgsql-performance-postgresql.org@postgresql.org>;
	Fri, 21 Sep 2007 08:33:57 -0300 (ADT)
Received: from postgresql.org ([200.46.204.71])
	by localhost (mx1.hub.org [200.46.204.182]) (amavisd-maia, port 10024)
	with ESMTP id 41602-03 for
	<pgsql-performance-postgresql.org@postgresql.org>;
	Fri, 21 Sep 2007 08:33:44 -0300 (ADT)
X-Greylist: from auto-whitelisted by SQLgrey-1.7.5
Received: from mail.ecircle.de (mail.ecircle.de [195.140.186.200])
	by postgresql.org (Postfix) with ESMTP id 95BCA9F98C7
	for <pgsql-performance@postgresql.org>;
	Fri, 21 Sep 2007 08:33:48 -0300 (ADT)
Received: from einstein.muc.ecircle.de (einstein.muc.ecircle.de [192.168.1.7])
	by mail.ecircle.de (READY) with ESMTP id 373CD55C002;
	Fri, 21 Sep 2007 13:33:46 +0200 (CEST)
Received: from [192.168.1.91] ([192.168.1.91]) by einstein.muc.ecircle.de with
	Microsoft SMTPSVC(6.0.3790.1830); Fri, 21 Sep 2007 13:33:45 +0200
Subject: Re: Linux mis-reporting memory
From: Csaba Nagy <nagy@ecircle-ag.com>
To: Heikki Linnakangas <heikki@enterprisedb.com>
Cc: Gregory Stark <stark@enterprisedb.com>, Decibel! <decibel@decibel.org>,
	pgsql-performance@postgresql.org
In-Reply-To: <46F39E4D.1080401@enterprisedb.com>
References: <20070920230400.GM95343@decibel.org>
	<16213.1190329934@sss.pgh.pa.us> <87r6ksa1jb.fsf@oxford.xeocode.com>
	<1190363401.4661.116.camel@PCD12478>
	<87r6ksuzf8.fsf@oxford.xeocode.com>
	<1190369325.4661.178.camel@PCD12478>
	<46F39E4D.1080401@enterprisedb.com>
Content-Type: text/plain
Date: Fri, 21 Sep 2007 13:34:22 +0200
Message-Id: <1190374462.4661.199.camel@PCD12478>
Mime-Version: 1.0
X-Mailer: Evolution 2.10.1 
Content-Transfer-Encoding: 7bit
X-OriginalArrivalTime: 21 Sep 2007 11:33:45.0862 (UTC)
	FILETIME=[4546BA60:01C7FC43]
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Archive-Number: 200709/358
X-Sequence-Number: 26903

On Fri, 2007-09-21 at 11:34 +0100, Heikki Linnakangas wrote:
> Which version of Postgres is this? In 8.3, a scan like that really won't
> suck it all into the shared buffer cache. For seq scans on tables larger
> than shared_buffers/4, it switches to the bulk read strategy, using only
>  a few buffers, and choosing the starting point with the scan
> synchronization facility.
> 
This was on 8.1.9 installed via apt-get on Debian 4.1.1-21. In any case
I'm pretty sure linux swaps shared buffers, as I always got worse
performance for shared buffers more than about 1/3 of the memory. But in
that case the output of top is misleading.

Cheers,
Csaba.