Mike,
we had similar problems a few months ago. The reason for TCP/IP failing
was a bug in the TCP/IP implementation for MVS:
When a client is idle for a too long time, the session gets cancelled,
that's how it should be. When the client is waiting for the host to
finish its database access, the host should - of course - not cancel the
connection. It seemed that the TCP/IP states got mixed up a little bit.

We got a bugfix PTF from IBM (I think V3.2) and the problem was gone.

Regards,
Michael
--
Michael H. Bartl
Michael H. Bartl
ADAC e.V. (Rechenzentrum, Produktionsplanung und Organisation)
Tel.: (089) 7676-4063, Fax: (089) 7676-8161
EMail: Michael.Bartl AT Zentrale.ADAC DOT de

> ----------
> Von:  mike murphy[SMTP:murphym AT FAATCRL.TC.FAA DOT GOV]
> Gesendet:     Mittwoch, 31. Dezember 1997 21:10
> An:   ADSM-L AT VM.MARIST DOT EDU
> Betreff:      tcpip timeouts during restores
>
> Hi,
>  We are running ADSM 2.1.0 on MVS.  We have two clients,  one UNIX and
> one
> NOVELL, trying to restore a large amount of files.  We are receiving
> message ANR0481W - session terminated - client did not respond within
> 120
> seconds.( these clients are doing their restores at different times so
> they
> are not conflicting with each other). We receive this error message
> after
> some files have been restored.  There does not seem to be excessive
> traffic
> on our network when this problem is occurring. Does anybody have any
> ideas
> what might be happening?  Thanks.
>