RE: Trashed disk

[Al Youngwerth <alberty@apexxtech.com>]

>From my SCSI days of long ago...

I believe your disk is telling you that it tried to write to a sector but that when it went to read it back to verify the data written was correct, the verify failed. Typically SCSI disks will then automatically reallocate the bad sector with some reserved sectors at the end of the disk. To find out for sure if the disk is saying this, you would need the tech manual for your fujitsu disk and look up the codes contained in the sense data.

If this is the case and you're consistently seeing these messages, I'd say your disk is spiraling downward. Eventually the disk will run out of sectors to re-map and a SCSI write will fail. If Linux does not support its own remapping on failed writes (does anyone know if it does?) you'll probably get a panic, your system will halt and you'll probably have a hell of a time trying to repair the volume and remount it to back it up.

Your disk could last for years to come or it could fail in minutes. My experience with these types of failures is that they tend to spiral relatively quickly (hours to days). I'd back up immidiately and start shopping for SCSI hard drives (almost free these days).

Al Youngwerth
alberty@apexxtech.com

----------
From: 	Bill Wohler[SMTP:wohler@newt.com]
Sent: 	Tuesday, September 24, 1996 1:30 AM
To: 	debian-user@pixar.com
Subject: 	Trashed disk

Folks,

  I've just noticed that I've been getting the following messages for
  the last couple of weeks.

Sep 24 00:00:28 gbr kernel: scsi0: MEDIUM ERROR on channel 0, id 2, lun
0, CDB: 0x03 00 00 00 10 00 
Sep 24 00:00:28 gbr kernel: Current error sr08:11: sns = f0  3
Sep 24 00:00:28 gbr kernel: ASC=10 ASCQ= 0
Sep 24 00:00:28 gbr kernel: Raw sense data:0xf0 0x00 0x03 0x00 0x06 0xc0
0xe9 0x28 0x00 0x00 0x00 0x00 0x10 0x00 0x00 0x80 
Sep 24 00:00:28 gbr kernel: scsidisk I/O error: dev 08:11, sector 442568

  My first reaction is to toss the disk, but I thought I'd see if anyone
  on this list would offer that it is salvageable (and if so, what
  precautions would be necessary).

  It's an old Fujitsu that I bought used 2.5 years ago and it's been
  running non-stop since.