Date: Wed, 4 Jun 2003 08:15:33 +1000
From: Nathan Scott <nathans@sgi.com>
To: xfs mailing list <linux-xfs@oss.sgi.com>
Subject: Re: xfs performance tuning
Message-ID: <20030604081533.C91882@wobbly.melbourne.sgi.com>
References: <3512.1054466831@ocs3.intra.ocs.com.au> <BAY2-DAV57lTbOzrPC40003086a@hotmail.com> <20030602230713.GC714@frodo> <BAY2-DAV29U29AhsFt4000049f7@hotmail.com> <20030603051946.GA1165@frodo> <20030603085416.GA7845@thomas.arkena.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5i
In-Reply-To: <20030603085416.GA7845@thomas.arkena.com>; from thomas@arkena.dk on Tue, Jun 03, 2003 at 10:54:16AM +0200
Sender: linux-xfs-bounce@oss.sgi.com
Errors-to: linux-xfs-bounce@oss.sgi.com
Precedence: bulk

On Tue, Jun 03, 2003 at 10:54:16AM +0200, Thomas Kirk wrote:
> Hep
> 
> On Tue, Jun 03, 2003 at 03:19:46PM +1000, Nathan Scott wrote:
> 
> > There are plenty more knobs to tweak, yes.  You should also
> > check that you are not doing any extra checks (e.g. do you have
> > debug or ACLs enabled?) that are not being done in ext2.
> 
> Does ACL decrase XFS performance considerable?
> 

Having ACLs enabled means slightly more work must be done, but
not much.  Having an ACL set on an inode means still more work
must be done, including additional disk IO when the ACL is not
being stored directly in the inode, so that can have an impact
on benchmarks.  Since ext2 wont be doing ACl stuff in the stock
2.4.20 kernels, it makes for a more direct comparison if ACLs
are not enabled in XFS too, of course.

cheers.

-- 
Nathan