We have been testing a TSM library manager. It is a TSM 6.2.2.0 instance
running under mainframe Linux. The tape drives we have been using for tests
are FCP-attached TS1120 drives in a 3494 library. A couple of weeks ago we
updated a drive to 'online=no' in our production environment and used the
drive for testing. We had no trouble defining the drive and the associated
paths on the library manager and running tape operations on the library
manager clients. We deleted the paths to the drive and the drive itself
from the library manager at the end of the day, but left the library and
the path to it defined. We tried a similar test today with two drives. We
were able to execute 'define drive' commands on the library manager, but
every attempt to define a path from the library manager to one of the
drives has failed with a message like:

ANR8420E DEFINE PATH: An I/O error occurred while accessing drive IBM7895753.

We have checked the command syntax carefully, and we have verified that itdt
(the successor to IBMtapeutil) can communicate with the drives. Messages
like the one above are the only diagnostic information in the activity log.
Executing 'set tapealertmsg on' does not change this. The TSM server code
obviously had information it did not reveal about which system call failed
and the nature of the failure. Is there any option or command which will
cause the server code to reveal this information?