Re: i2o_scsi w/ tape and RAID kernel errors

From: vyh (vyh.1xlpih_at_no-mx.usenetlinux.com)
Date: 10/28/05 

Date: Fri, 28 Oct 2005 02:02:47 -0500

HI,
I've been having a similiar problem on a system that has recently been
upgraded to RHEL4. The tapelibrary is an IBM ultrium and although it
doesn't lock the server, it constantly loses the device - ie, at least
one tape drive goes off line, often locking up that drive until I reset
it.

dmesg: tapelib kernel: scsi-osm: SCSI error 05000002

errors appeared after I started accessing the device.

before that from messages:

Oct 25 16:31:58 tapelib kernel: scsi1 : Adaptec AIC79XX PCI-X SCSI HBA
DRIVER, Rev 1.3.11
Oct 25 16:32:02 tapelib kernel: Attached scsi generic sg0 at scsi0,
channel 0, id 0, lun 0, type 0
Oct 25 16:32:02 tapelib kernel: Attached scsi generic sg1 at scsi0,
channel 0, id 1, lun 0, type 0
Oct 25 16:32:02 tapelib kernel: Attached scsi generic sg2 at scsi0,
channel 0, id 2, lun 0, type 0
Oct 25 16:32:02 tapelib kernel: Attached scsi generic sg3 at scsi0,
channel 0, id 3, lun 0, type 0
Oct 25 16:32:03 tapelib kernel: Attached scsi generic sg4 at scsi0,
channel 0, id 4, lun 0, type 0
Oct 25 16:32:03 tapelib kernel: Attached scsi generic sg5 at scsi0,
channel 0, id 5, lun 0, type 0
Oct 25 16:32:03 tapelib kernel: Attached scsi generic sg6 at scsi0,
channel 0, id 8, lun 0, type 3
Oct 25 16:36:29 tapelib kernel: scsi2 : iop0
Oct 25 16:36:29 tapelib kernel: Attached scsi generic sg7 at scsi2,
channel 1, id 4, lun 0, type 1
Oct 25 16:36:29 tapelib kernel: Attached scsi generic sg8 at scsi2,
channel 1, id 5, lun 0, type 1
Oct 25 16:36:29 tapelib scsi.agent[2695]: tape at
/devices/pci0000:01/0000:01:0c.1/host2/target2:1:5/2:1:5:0
Oct 25 16:36:29 tapelib scsi.agent[2678]: tape at
/devices/pci0000:01/0000:01:0c.1/host2/target2:1:4/2:1:4:0
Oct 25 16:36:29 tapelib kernel: Attached scsi tape st0 at scsi2,
channel 1, id 4, lun 0
Oct 25 16:36:29 tapelib kernel: Attached scsi tape st1 at scsi2,
channel 1, id 5, lun 0
Oct 25 16:36:29 tapelib kernel: Attached scsi generic sg9 at scsi2,
channel 0, id 6, lun 0, type 8
Oct 25 16:36:29 tapelib scsi.agent[2924]: changer at
/devices/pci0000:01/0000:01:0c.1/host2/target2:0:6/2:0:6:0

So its apparent that the two adapter cards are being seen and the
devices are loading.

I'm not sure what to do from here - the I2O pages suggest that you
shouldn't run scsi_i2o for raid devices (incidently, I'm running
software raid - the physical machine is a Fujitsu-Siemens server and
their raid software refuses to recognise the devices in a consistant
manner.) Another post suggested adding a patch (I HAVE to run
the 2.6.9-11.EL kernel, as the IBM tape drivers are kernel specific.)

I would appreciate a reply from anyone who has any clues to where to
best go from here, as this is our backup server, and therefore rather
critical.

Cheers,
Vera :) 

-- 
vyh
------------------------------------------------------------------------
vyh's Profile: http://usenetlinux.com/member.php?userid=5635
View this thread: http://usenetlinux.com/showthread.php?t=129797