On Wednesday, 11/07/2007 at 01:16 EST, Thomas Kern <[EMAIL PROTECTED]> wrote: > Does 'FORCE RSCS' quiesce all non-active links and wait (at least for the > shutdown-defined within period) so that the MVS systems don't try to send > new files while RSCS waits for that large printout to finish on that slow > network-attached printer?
If you FORCE RSCS, RSCS won't be waiting for much of anything, will it? It won't sign-off, true, and MVS should go into error recovery trying to reestablish a connection if they send a file. > Does 'FORCE TCPIP' bring down all of the TCPIP related servers that TCPIP > started up? (I hope that the ~CP EXT entered on the TCPIP console will do > that now, yes, # can be disabled as linend and if you like ~ put in its place.) No. I think the NFS server waits for the stack to come back. But just force them off, too. After all, when the stack comes back up, it will force/autolog them, so what's the diff? > There are SVMs that may have outstanding work to do when a system has been > requested to shutdown. Can you imagine the problems if DATAMOVE were FORCEd > in the middle of moving Chuckie's 191 minidisk? I was under the impression that the DATAMOVE request would remain in queue and be performed again when DATAMOVE came back up. But I'm willing to be educated on this point. However it works, it's been that way for decades and we haven't (to my knowledge) received any Integrity APARs on it. > Once you have set out on the road to cleanly shutdown SVMs that may have > outstanding work to do (Linux and SFS), why stop there, why not clean up > your older SVMs too? You can, but do you *need* to? Alan Altmark z/VM Development IBM Endicott
