> >I'm not sure that you're aren't describing a bug somewhere, but is it > >technical problem that you say can be reasonably avoided by orderly > >shutdown. So I'd say: write up an RSCS requirement to support signal > >shutdown. > > It may be a bug, but it's a persistent bug. We've had the "problem" > for years. Therapy helped, as did banging our heads on the concrete wall > outside. We finally settled on Netview automation on the MVS side to > look for the problem, drain and restart the link. A lot less blood on the > floor.
Might be related to which side of the connection wants to be master. JES tends to want to be the controlling side, and it may take a few restarts to ensure that the VM side becomes properly submissive during the initial link setup negotiation. JES3 *must* be the master (it won't negotiate), so you get this a lot there, but JES2 doesn't like being slave, and will try *really hard* not to end up in that role.