[4421] in linux-scsi channel archive
Re: Bad blocks: How many?
daemon@ATHENA.MIT.EDU (Dr. Michael Weller)
Tue Jul 21 09:02:45 1998
Date: Tue, 21 Jul 1998 12:14:52 +0200 (MESZ)
From: "Dr. Michael Weller" <eowmob@exp-math.uni-essen.de>
To: Tim Robinson <timr@sykes.demon.co.uk>
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <01bdb48c$96b04f00$0102a8c0@gateway.melksham>
On Tue, 21 Jul 1998, Tim Robinson wrote:
>
> Hi
>
> I lately discovered scsiinfo from this list, which I ran on
> my IBM DCAS 34330 (4.33GB) drive.
>
> scsiinfo claims 647 defects in the manufacturers list, then throws out a
> list,
> finally reporting '0 entries in grown table'
>
> Is 647 a reasonable defect count? I have no feel for what to expect here.
Well, while I (and probably you too) personally claim anything above 0
defects unreasonable, the pressure of price and stuff imposes other
values. I've seen other disks (not IBM too) with defects in that range
which did not result in problems. Also, in which format is this list? is
it 647 Sektors bad or 647 Tracks bad (the first won't be a big deal, but I
assume it is the second).
> This disk has been giving occasional problems, causing scsi timeouts and
> preventing some users from saving large files. Is this related?
Well, at first there are 0 grown defects, so what. Of course, you should
check the AWRE and ARRE settings and the amount of spare sectors to use in
case of defects. If the disk cannot reallocate bad sectors due to wrong
settings there, they won't show up as defects. (actually I'm not 100% sure
if such on the fly defects are added to the grwon list or if you have to
run a format and disk certify process to add to the grown list).
Anyway, in case of any errors here, I won't expect timeouts. I assume the
timeouts are either to "EXTREMELY BROKEN" disk (bad heads sticking to
surface, bad motor with wrong spin speed, whatever; not too likely) or
(much more likely) either 'SCSI disconnection' not enabled or not
understood by all connected devices (s.t. a slow CDROM, TAPE, other disk..
blocks the SCSI bus resulting in other outstanding requests to timeout)
or cabling problems resulting in commands/replies to get lost.
Note that the pure existence of a SCSI-2 or ULTRA (or whatever) device on
the bus requires that the cabling limits of this specs are met. For
example, lets assume the disk is SCSI-ULTRA (is tnat 40MHz bus speed?
anyway..) but your controller does only Fast-SCSI (something like 5 or
10).. Well still, you must use cables approved for ULTRA and active
termination nevertheless, just because the input lines of the new and fast
disk are soo sensitive regardless of what the controller actually
supports.
I'd assume a problem in this direction.
Michael.
--
Michael Weller: eowmob@exp-math.uni-essen.de, eowmob@ms.exp-math.uni-essen.de,
or even mat42b@spi.power.uni-essen.de. If you encounter an eowmob account on
any machine in the net, it's very likely it's me.
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
