OK, I've been chasing this one for weeks and I'm at my Witt's End (and
already read the magazine and recovered the pearl)...

We had DTC's that stopped responding (users complained, serial
printer spoolers stopped, etc) several times a day.  From the front
panel display, they were either attempting an upload or reporting
failure of upload (DTC Manager based).  Either power-cycling the DTC
or resetting it from DTC Mgr would reload it and it would go on about
it's merry way for awhile.

This was not one DTC, but arbitrarily any of ours with any remaining
significant use (out of 6, I'm sure 5 have failed at least once, and
I vaguely recall the other one dying too); thus I ruled out anything
hardware.  It smelled like either a network problem or perhaps some
hacking attack of a DoS (Denial of Service) nature, but I've never heard
of any typical hacks bothering a DTC before.  As to the network,
there were no significant errors reported on the DTSLINK interface,
nor on the switch port where the segment attached to the campus LAN.
Nothing has changed in years in the topology or equipment.

Network-wise, it was a ThinLAN segment running from our 3000/969, across
some vacant MAUs leftover from the 950, chaining through the
DTCs, then running some more vacated cables/T-connectors, and finally
going to a 10Base2-to-10BaseT repeater, and finally into the switch.
Suspecting something in the dead real estate, after a week or two of
these strange errors Richard and I rewired the works, eliminating
everything "dead", replacing the repeater and most of the cable
segments.  No luck eliminating the problem.

As a final act of desperation, I converted the DTCs from ThinLAN to
the AUI port, attached SQE-enabled transceivers, and ran cables directly
to our central Cisco switch (as direct a path as possible).
Yet today, the problem is still occurring.  Aaaargh!

I did however get some completed DTC uploads, and have a call in to
the HPRC to get them analyzed (aggravating that they don't give you
ANY tools to get even basic information from these).

Does anyone have any insight or past horror stories like this one that
might shed some light on the problem?

Jeff Kell <[log in to unmask]>