SCS_AUTO_SENSE_FAILED Problem with HVD SCSI card in Octane.

[m***@gmail.com]

Recently we had a SGI qlogic HVD SCSI card installed in a older SGI
Ocatne and attached to a Sony DTF2 tape deck model #

The card had worked months/years in the machine with no issue.

Now for no reason I am getting SCS_AUTO_SENSE_FAILED errors upon boot
and the drive no longer functions on the machine.

I have swapped everything except the differentail termainator we have
used for year and the problem occurs with two different tape drives,
two different Octanes, two different HVD SCSI cards, two different I/O
card cages and numerous different cables

here is the output from the SYSLOG

May 24 15:10:28 6A:chicago unix: ql3d4: <6>HARD ERROR: completion
status
[0x1e] (SCS_AUTO_SENSE_FAILED), scsi status[0x2], state flags[0x1f0
0], status flags[0x84]
May 24 15:10:28 5B:chicago MAKEDEV_tape: Cannot find any tape device;
link to tape not created.
May 24 15:10:28 6D:chicago sesdaemon[569]: |$(0x339) No fibre channel
controllers found
May 24 15:10:29 1D:chicago mediad: Can't dsopen /dev/scsi/sc3d4l0: No
such device
May 24 15:10:29 6A:chicago unix: ql3d4: <6>HARD ERROR: completion
status
[0x1e] (SCS_AUTO_SENSE_FAILED), scsi status[0x2], state flags[0x1f0
0], status flags[0x84]
May 24 15:10:30 6A:chicago unix: ql3d4: <6>HARD ERROR: completion
status
[0x1e] (SCS_AUTO_SENSE_FAILED), scsi status[0x2], state flags[0x1f0
0], status flags[0x84]
May 24 15:10:30 1D:chicago mediad: can't open tape device
/dev/rmt/tps3d4stat: No such device
May 24 15:10:30 6A:chicago unix: ql3d4: <6>HARD ERROR: completion
status
[0x1e] (SCS_AUTO_SENSE_FAILED), scsi status[0x2], state flags[0x1f0
0], status flags[0x84]

I am kind of at wits end

[Simon Burley]

Post by m***@gmail.com
Recently we had a SGI qlogic HVD SCSI card installed in a older SGI
Ocatne and attached to a Sony DTF2 tape deck model #
The card had worked months/years in the machine with no issue.
Now for no reason I am getting SCS_AUTO_SENSE_FAILED errors upon boot
and the drive no longer functions on the machine.
I have swapped everything except the differentail termainator we have
used for year and the problem occurs with two different tape drives,
two different Octanes, two different HVD SCSI cards, two different I/O
card cages and numerous different cables

If you have two of everything (drives, card cages, cables, HBAs,
Octanes) then presumably you have two complete, identical systems?

The strange thing is that IRIX has probed the bus and a device has
answered at address 4. I think that it is cabling or terminator failure.
Maybe a reflection on the bus.

Can you do a uname -aR and let us know which IRIX version is on the
host? Are you totally sure that nothing changed on the host?

If you were in South London you'd be welcome to try swapping parts with
out DTFs. Can you find someone locally with, say, a differential disk
drive to test the bus with?

I think the qlogic boards are not all that great. Certainly they don't
negotiate/honour disconnects.