[1236] in linux-scsi channel archive
Re: Update (Re: medium error, what next?)
daemon@ATHENA.MIT.EDU (Bernd =?UNKNOWN-8BIT?Q?Anh=E4upl?=)
Wed Jan 8 01:39:19 1997
Date: Tue, 7 Jan 1997 16:44:45 +0100
From: Bernd =?UNKNOWN-8BIT?Q?Anh=E4upl?= <anhaeupl@late.e-technik.uni-erlangen.de>
To: srb@cuci.nl
CC: linux-scsi@vger.rutgers.edu, ncr53c810@Colorado.EDU
In-reply-to: <199612291427.PAA01408@dionysus.cuci.nl> (srb@cuci.nl)
From: srb@cuci.nl (Stephen R. van den Berg)
Date: Sun, 29 Dec 1996 15:27:35 +0100
Sender: owner-ncr53c810@mroe.cs.colorado.edu
Precedence: bulk
Well, I thought I'd wait a little to see if the drive stays stable...
But I think I can now say that it looks pretty good.
This is the drive:
ncr53c810-0-<target 3, lun 0>: using tagged command queueing, up to 4 cmds/lun
Vendor: IBM Model: DPES-31080 Rev: S31Q
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdd at scsi0, channel 0, id 3, lun 0
Strangely scsiinfo shows that AWRE and ARRE were both on (I didn't change
any of the drive parameters). Nonetheless, simply reading the entire
drive did not result in the faulty sectors being relocated.
(BTW, I still don't know what AWRE and ARRE exactly enable and/or disable).
Simply writing the relevant sectors (two spots, eventually) made the error
go away (I simply wrote through the buffer system on Linux, then sync
afterward). The strange thing is that the drive doesn't show any
errors in the "Grown defect list"; which makes me wonder if it relocated
anything at all? Maybe it just rewrote the sector and in doing that, it
wiped out the error that was (perhaps) due to some weak magnetisation
in that spot.
...
Did you switch of the power of your computer, when your drive has (or might have)
been in the middle of writing a sector sometime in the past?
We got exactly one bad sector on a SCSI harddisk connected to an HP workstation this
way. In this case the drive has really __no__ chance to read that sector correctly,
since its first part probably contains new data, followed by some corrupt data and
the rest of the sector should still contain old data. Therefore the drive
cannot relocate the sector, since it won't be able to read that sector, and there
is also no need to relocate that sector (only to rewrite it), since probably the disk
is still ok at that place.
After rewriting the sector with a little C program, that drive has been running
now without any problems for at least about 2 years.
--
Bernd Anhaeupl Tel.: +49 9131 857787
LATE - Uni Erlangen
Cauerstr. 7 Email: anhaeupl@late.e-technik.uni-erlangen.de
91058 Erlangen
