Subj : DDS4 Tape Errors
To   : comp.os.linux,comp.periphs.scsi
From : nickbarendt
Date : Wed Jun 23 2004 09:12 am

I have a Dell PowerEdge 2600 running RH9 (2.4.20-30.9) with a PERC
4/Di "MegaRaid" RAID/SCSI controller.  There is a Dell PowerVault 100T
DDS4 (aka Archive Python 06408, firmware rev 9100) SCSI tape drive on
the secondary SCSI channel.  The tape drive is the only device on the
SCSI channel and there is a terminator on the end of the cable and the
controller's termination is enabled on the other end.
  The system has been running for about 8 months with no trouble,
using Amanda to perform daily backups.  Recently we started having
some tape drive trouble, though.  Initial problem was that some error
during the backup would lock up the entire machine - everything.  Dell
technical support suggested forcing the SCSI channel speed down to
Fast from the default 320 MB/s.  That seems to have fixed the the
lockup problem, but then we started seeing other drive errors.  Dell
swapped the tape drive, cable, and terminator under warranty (Yay
Dell!), but the errors have continued.  They seem to be intermittent. 
Some times the backup runs fine, other times it fails with errors like
this:

Jun 19 02:23:43 kernel: st0: Error 250400f0 (sugg. bt 0x20, driver bt
0x5, host bt 0x4).
Jun 19 02:23:44 kernel: st0: Error with sense data: Current st09:00:
sense key Unit Attention
Jun 19 02:23:44 kernel: Additional sense indicates Power on,reset,or
bus
device reset occurred
Jun 19 02:23:44 kernel: st0: Error on write filemark.

I haven't been able to decode the Error code of 250400f0.  The backup
is failing relatively early, somewhere less than 3GB of an 18GB
backup.

We keep up with the cleaning tapes and the tapes in the regular
rotation are of varying ages, 8-4 months old, and I have not seen any
correlation of errors with tape age.

The size of the backup has increased slightly in recent weeks, but not
dramatically, going from about 16GB to about 18GB.  The drive's
hardware compression is disabled - Amanda is configured to use
software compression for better tape usage estimation.

The latest from Dell Tech support was to update the firmware on the
RAID/SCSI controller, now at rev 2.48.  That does not seem to have
remedied the problem -we're still getting the same errors.

I would appreciate any hints or tips that anyone can offer.

Nick Barendt
VXI Technology, Inc.

.