[870] in linux-scsi channel archive
Re: scsi 2940 controllers
daemon@ATHENA.MIT.EDU (Gerard Roudier)
Sun Nov 3 12:42:49 1996
Date: Sun, 3 Nov 1996 19:40:06 +0000 (GMT)
From: Gerard Roudier <groudier@club-internet.fr>
To: Matthias Urlichs <smurf@smurf.noris.de>
cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <327c81c9.2658-234.2@work.smurf.noris.de>
On Sun, 3 Nov 1996, Matthias Urlichs wrote:
> In linux.dev.scsi, article <2.2.32.19961028192441.00693f04@mail.bluesky.net.au>,
> tommiy <tommiy@bluesky.net.au> writes:
> > Can anybody please help stop my system from locking up, please. It seems to
> > be an impossible task.
> >
> In fact, it is very easy to do that..:
>
> > heading because I just cant get linux to be stable! Can someone else offer
> > some suggestions as to what causes the error below?
> >
> > scsi0: MEDIUM ERROR on channel 0, id 3, lun 0, CDB: 0x0 a 13 29 c1 f4 00
>
> Did you ever look at the error message, and think about what MEDIUM ERROR
> might actually mean?
>
> You just have a bad spot on your hard disk.
>
> Now, the fact that the system doesn't recover from the error _is_ a bug, all
> right, but it's not the source for your problem...
Agreed about the source of the problem, but not about the _Linux_ bug.
For example:
- Unrecoverable read error cannot be recovered by the firmware until you
do a destructive write to the spot with AWRE enabled (and obviously
supported).
Remapping the sector with some famous disk tool is impossible without
data lost.
Only user is responsible of lost data.
In the initial post, it was written:
ASC=10 ASCQ=00 that should mean "ID CRC OR ECC ERROR".
I am not familiar with such error, but it seems to me unrecoverable even
with a ZeroBugSystem(TM).
(Ok, it is a plagiarism. Sorry, but I could'nt resist)
Gerard.
