Message-ID: <3B044CEE.16BEE674@cfl.rr.com>
Date: Thu, 17 May 2001 18:13:02 -0400
From: Mark Hounschell <dmarkh@cfl.rr.com>
Reply-To: dmarkh@cfl.rr.com
MIME-Version: 1.0
To: Seth Mos <knuffie@xs4all.nl>
CC: kris buggenhout <buggenkr@god.bel.alcatel.be>,
   "linux-xfs@oss.sgi.com" <linux-xfs@oss.sgi.com>
Subject: Re: trouble benchmarking
References: <4.3.2.7.2.20010517130519.038d0c88@pop.xs4all.nl>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-linux-xfs@oss.sgi.com
Precedence: bulk

Seth Mos wrote:
> 
> At 11:46 17-5-2001 +0200, kris buggenhout wrote:
> >it looks like a scsi error .. but when I do the same test on the same
> >drive with ext2 i get no error.
> 
> ext2 probably handles bad blocks better. The XFS filesystem shuts down on
> error. Ext2 may stay up and note that the sector is moved and the data is ok.
> It might be possible that xfs could be made to respect this re allocation
> of a bad sector and perhaps not shut down.

The scsi drives "page-1 mode-select" error recovery page should be set
to not
report recovered errors. This can be changed with any decent scsi disc
utility
and does not require and reformat or lost data after the change. No
body's driver/
filesystem should have to worry about that. You shouldn't see the
message again
unless youv'e got more bad spots and don't change page-1 to not post
recovered
errors.

Regards

-- 
Mark Hounschell
dmarkh@cfl.rr.com