Date: Wed, 31 Aug 2005 21:22:17 +0200
From: Tobias Brox <tobias@nordicbet.com>
To: Carlos Henrique Reimer <carlosreimer@yahoo.com.br>
Cc: pgsql-performance@postgresql.org
Subject: Re: Swapping
Message-ID: <20050831192217.GC5736@tobias.lan>
References: <20050831182515.89931.qmail@web53209.mail.yahoo.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
In-Reply-To: <20050831182515.89931.qmail@web53209.mail.yahoo.com>
Organization: Group Nordicbet
User-Agent: Mutt/1.5.10i
Content-Transfer-Encoding: quoted-printable

[Carlos Henrique Reimer - Wed at 03:25:15PM -0300]
> I=B4m trying to tune a linux box with a 12 GB database and 4 GB RAM. Fi=
rst
> of all I would like to stop the swapping, so the shared_buffers and sor=
t_mem
> were decreased but even so it started swapping two hours after DBMS sta=
rted
> up.
> =20
> I would like to know some suggestions how to discover why is it swappin=
g?

I agree with Tom Lane, nothing to worry about.  Swapping is not a problem
per se, aggressive swapping is a problem.  If you are absolutely sure you
want to ban all swapping, use "swapoff"?

I'd trust linux to handle swap/cache sensibly.  Eventually, become involv=
ed
with kernel hacking ;-)

--=20
Notice of Confidentiality: This email is sent unencrypted over the networ=
k,
and may be stored on several email servers; it can be read by third parti=
es
as easy as a postcard.  Do not rely on email for confidential information=
.