[plug] Urgent RAID Help needed

[Grahame Bowland]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 31 Jul 2002 07:21 pm, Matt Kemner wrote:
> On Wed, 31 Jul 2002, quoth James Bromberger:
> > Anyways, checking the log showed:
> > > Jul 30 03:03:27 phobe kernel: hdc: dma_intr: status=0x51 { DriveReady
> > > SeekComplete Error }
> >
> > two disks set into a group of identicla partitions. Now, if a disk was
> > bad, I would have expected all partitions on that disk to have gone bad,
> > not just partition 7 (which, of course, has may main fileshare on it, of
> > course).
>
> Not necessarily
>
> That error usually means "bad sectors" - which would only affect a small
> area of one partition, probably limited to a handful of files. In fact,
>
> I've got a drive here which I've been using for storage of large (600MB+)
> files, which has bad sectors on it, and all I did was move the offending
> files to a /DAMAGED/ directory, and the rest of the drive was fine for a
> while.  However, once you start getting bad sectors, they tend to spread,
> so now would be a good time to return that drive of yours for warranty.
> Mine's >5 years old, so I'm afraid that's not an option.

Actually, I've had similar experiences to James in the past. One of my servers 
spontaneously decided to fail half of the disks in the box; these disks all 
corresponded to one of the SCSI buses. It would only happen intermittently; 
turned out to be the SCSI cable.

Just goes to remind you that you're definitely best off putting one half of 
the RAID set on one bus, the other half on the other; this also has obvious 
performance benefits.

- -- 
Grahame Bowland - <grahame at azale.net>
'All programmers are playwrights and all computers are lousy actors'
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE9R8mfN7eIygKWKbARAvr8AJwMUtd8WhuyqxZVF6/PZVVqbq2iHwCfY8QF
tZjG2/yOcA08clI8DLfyziE=
=vooi
-----END PGP SIGNATURE-----