On Tuesday, 11/03/2009 at 04:45 EST, Bob Levad <ble...@winnebagoind.com> wrote: > I don't know if others have seen this behaviour, but I've seen discussion of > logoff/force pending recently and found nothing similar in a quick internet > search.
If you LOGOFF or are FORCEd while you are in the middle of LOGON HERE, you should get HCPLGA6051E Restart the logon procedure because reconnect processing cannot be done Any time you get LOGOFF/FORCE PENDING and the condition doesn't clear itself in a reasonable[1] amount of time you should get a SNAPDUMP or restart dump and open a PMR. If CP isn't going to give you a hint as to what to do to fix the problem (other than IPL), then CP needs to not get into that state in the first place. Hence the need to open a PMR. Alan Altmark z/VM Development IBM Endicott [1] The longest you have to wait for an I/O to complete or get an error is twice the missing interrupt handler time for the device. INDICATE I/O will tell you if the user is waiting on I/O. Do a QUERY MITIME on any device you see the user waiting on. Take the longest MITIME and double it. For DASD and tape, the MITIME value comes from the device itself. Note that for 2nd level systems, the missing interrupt handler is turned OFF by default!