On Nov 8, 2005, at 3:09 AM, Alexander Verkooijen wrote:

We also have a number of Windows 2000 clients
(5.3.2.0) that seem to hang the entire system
during a backup. The client has to be rebooted
to get it working again. ...

The poster of the original question had a TSM server which was
utilizing default timeout values - which are far too low for real
world conditions, where clients have to spend a lot of time rummaging
around in file systems seeking backup candidates before they next
interact with the server. The server may deem such sessions hopeless
and terminate them from its end, where sometimes the client "doesn't
get the message" that the session was cut off.

In all "hang" situations, you need to dig in and extract the details:
most "hang" situations are not hangs at all, but peer situations
where one side is busy and the other is waiting for it. Look for ANR
messages in your TSM server activity log, any OS or dsmerror.log
entries on the client, monitor process activity, I/O activity, etc.
to ascertain what's really going on. You *may* have a situation that
an APAR may cure, but you first need to identify symptoms so as to
apply the right remedy. A W2K system which is downlevel in service
packs can be the cause of their own problems.

   Richard Sims