Date: Wed, 9 Apr 2008 10:00:45 -0400
From: Andrew Sullivan <ajs@crankycanuck.ca>
To: pgsql-admin@postgresql.org
Cc: Johann Spies <jspies@sun.ac.za>
Subject: Re: Handling large volumes of data
Message-ID: <20080409140045.GA14729@crankycanuck.ca>
Mail-Followup-To: Andrew Sullivan <ajs@crankycanuck.ca>,
	pgsql-admin@postgresql.org, Johann Spies <jspies@sun.ac.za>
References: <20080408094233.GG13409@sun.ac.za> <200804081201.15175@zmi.at>
	<20080408101336.GJ13409@sun.ac.za>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20080408101336.GJ13409@sun.ac.za>
User-Agent: Mutt/1.5.13 (2006-08-11)

On Tue, Apr 08, 2008 at 12:13:36PM +0200, Johann Spies wrote:
> I have got 8x720G disks in a hardware raid 5 setup. It is a Dell 2950

Thow away your RAID 5.  It's a loser for this.  Raid 1+0 is what you need.

> server. I am using an XFS-filesystem.  I am not certain about the

On another note, I've had abysmal experiences with xfs on linux.  Like,
"Oops, everything's in lost+found now!" abysmal.

(Unfortunately, actually, in my experience Linux has only bad and worse
choices for database filesystems.  Maybe they'll get it right with ext4.)

A