And especially with CMS, testing old stuff on new CMS on old CP usually presents no problems. We did have to fall back due to a problem in OEM CP modifications, so we are running 4.4.0 with nearly 90 users on CMS 22. We are not using VMSYS and have had no problem with the filepool servers that we are running on CMS 22. Are you sure that your VMSYS had the correct 193 disk when it failed? As long as that disk is correct for the release of CMS, running CMS 22 on z/VM 4.4.0 appears to work as it should. Mixing a back level 193 disk with a newer CMS does not work.
Regards, Richard Schuh -----Original Message----- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Alan Altmark Sent: Monday, May 08, 2006 7:09 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: CMS Levels On Monday, 05/08/2006 at 09:05 AST, "Stracka, James (GTI)" <[EMAIL PROTECTED]> wrote: > We had a problem in the past when running an SFS SVM at a higher CMS > level than CP. Specifically the VMSYS: SVM took an abend. We now > insure that CP is always higher than or equal to both CMS and GCS. Generally the order to use is: - CP first - then CMS (with LE) - then TCP/IP (stack, apps, SSL) - GCS whenever >From time to time there may be specific exceptions that need tigher bindings that you usually (!) hear about in advance, but this generally works. If you wander too far afield from this order, you risk getting yourself into an unsupportable configuration. We don't test new-on-top-of-old, but typically test old-on-new as that reflects the realities of our development and incremental "function test" process. The full-blown "system test", however, is primarily a new-on-new environment since it proceeds according to the documented procedures for installing a z/VM system. Alan Altmark z/VM Development IBM Endicott