Richard,
        We are experiencing something similar to this, but like another 
response earlier, we are not seeing the hang condition on the Library Manager. 
We do see drives with the tape stuck in it, and most often we are able to 
remove it by simply selecting the unload option on the drive menu(which says 
to me that the error wasn't a permanent hardware failure).  One item to note, 
we recently installed a new level of "microcode" on our Library Manager(the 
PS/2 system running the Library Manager program).  This happened about a month 
ago, and we have been having problems for at least 2 weeks now.  This seems to 
me to be quite a coincedence.
        You mention that it is a communication problem, which I have been 
suspecting for a while now, due to the fact that I have noticed ANR8824 
error messages in the activity log as of late.  An example:

10/04/99   06:28:10  ANR8824E I/O Error on library L3494; request 0F0DF5AE for 
                      operation 004C6D31 to the 3494 Library Manager been lost.
10/05/99   05:55:49  ANR8824E I/O Error on library L3494; request 0F0DFFC0 for 
                      operation 004C6D31 to the 3494 Library Manager been lost.

These errors combined with the simple manual unload of the drive, coupled with 
the new level of code on the Library Manager, seem to point to a bug in the 
Library Manager.  Does this sound logical?  Where can I get an updated version 
of the LM "microcode".  I don't really see anything in the RShelp databases, 
hopefully the CE can get a hold of it.

-Matt
>A note from the field as to what can happen with a 3494...
>A note from the field as to what can happen with a 3494...
>I came in yesterday morning to find an Int Req on the robot:
>both 3590 drives in the last frame had failed to unload, at very
>different times of the night.  The two drives in the first frame
>had no problems.  After some 12 hours of analysis and erroneous
>guesses by various IBM levels, it was finally determined that the
>ARTIC (A Real-Time Interface Coprocessor) had partially failed.
>This card in the industrial computer within the 3494 manages
>RS-232 and RS-422 communication, as serial connections to a host
>and command/feedback info to the tape drives.  The last two drives
>had ejected their tapes and had tried to tell the Library Manager
>about this, but could not.  When the robot needed to re-use the
>drive but knew there was a tape in it, it posted the Int Req.
>Attempting to clear the Int Req at the control screen resulted in
>the LM being hung with a clock icon.  A Ctrl-Alt-Del reboot would
>then always result in a memory dump and reversion to a Shutdown
>selection box.  With all the 3494's internal logging one would
>think that diagnosis of such a problem would be straightforward;
>or at least the manifested symptoms would be well-known.  Not.
>
>So the next time you get an Int Req and the tape is sticking out
>of a drive but not removed by the robot, suspect a problem with
>the drive's communication with the library manager.
>      Richard Sims, BU


*---------------------------------------------------------*
Matt Bacchi                              mbacchi AT us.ibm DOT com
IBM Global Services                    v2bacchi AT btv.ibm DOT com
D54V Server Infrastructure
ADSM & AFS/DFS Backup