The library mechanism is having difficulty communicating with the drive

[yeok song]

Hi,

this is the second time i facing this problem. Everytime i need to restart the tape drive to solve the problem. when this problem occur the scsi bus will reset. Anybody know the cause? I'm using DLT autoloader with omniback II 3.5.

Thanks.

[Michael Tully]

Hi,

There are a few things I would like
to suggest to you. DLT drives and autoloaders
do not like receiving SCSI resets at all.
You need to switch off the EMS monitoring for
SCSI tapes on your server. I have attached
a document that should help with doing this
task. Second thing is have a look your firmware
level for your autoloader.

Also have a look at the OmniBack patches.

HTH
-Michael

[David Ruska]

A few comments on EMS: First, EMS only needs to be disabled (or alternatively have polling disabled) IF the tape devices are in a SAN environment, where multiple hosts can access the tape drives (through the SAN). In this case, the easiest solution is turn off polling (by setting POLLING_INTERVAL to 0 in the dm_stape.cfg file) if the EMS version is September 2001 or newer (IPR0109).

Second, the instructions provided are outdated, and will actually cause an additional problem of logtool failing to decode SCSI logs. The step stating to rename dm_stape [mv /usr/sbin/stm/uut/bin/tools/monitor/dm_stape
/usr/sbin/stm/uut/bin/tools/monitor/dm_stape.orig] should NOT be done. This monitor is also used as a decoder for the stape driver.

Regarding the original problem, more information on the product, configuration, error logs, etc would be needed to give you a good diagnoses.