Hi guys,

First I want to thank all the people who wrote to me on the
subject. I think I found the problem.

It became clear that it was a timing problem. Furthermore, I was
puzzled by the fact that if I had the machine off for a few
hours, that means everything cold it would work. As soon as the
thing got warmed up, the failure appeared. That seems to
indicate that could be related to the fact that the drivers can
deliver more "juice" when they are cold. My SCSI configuration
was: two CD-rom drives (a Chinon and an HP), the TK50 and the
disk (a Seagate Barracuda). So I unplugged the Chinon thinking
that if I reduce the loading it may work... and Magic! It did!
The Chinon was perfectly functional, but it seems that it was
loading the bus too much. So I replaced that Chinon by another
HP drive, which doesn't seem to load the bus. So everything
works fine now.  We could say "not all SCSI devices are created
equal"...

Dan

>> I'm having problems with my TK50 tape drive. Most of the times
>> when I read or write on it it crashes the system with the
>> following console messages (approximately because I copied them
>> by hand):
>> 
>> asc_intr: tgt 5 command phase TC zero with non-empty ...
>> asc_intr: tgt 5 command phase TC zero; padding command
>> asc_get_status: cmd reg 11, fifo cnt 3
>> 
>> and/or:
>> 
>> asc_mtr: 1: len 249 (fifo 0)
>> asc0: SCSI device 5: rejecting message 0x0
>> asc0: SCSI device 5: unexpected disconnect
>> 
>> I try many combinations of hardware changes:swaping TK50 drives
>> and cables, etc. Can somebody help? Thanks in advance.
>
>I get the same sort of errors.  My fix, was to bring up a box on Ultrix
>to write TK50 tapes.  This is on pmaxen and vaxen on 1.4P.  It seems worse
>on 5000/200 class machines compared to lesser 5000 or 3000 or 2000 machines.
>It seems to happen on any vaxen of the 3100 class.
>
>I also get asc_intr errors on HD accesses occasionally, but they seem to
>be more spurious interrupt things.
>
>I sense is is due to timing problems perhaps, since it does not seem
>to occur on slower machines nearly as often.
>
>TK 50 writes and reads just don't work, but the HD errors don't seem to
>do any harm.
>
>Anyone else seen this sort of asc_intr problem?  Any fixes?