I don't know if this will help. We are not running z/OS under z/VM. The MIH for all of our DASD is set to two minutes (DASD=02:00). Here is the JES2 info on the checkpoint. DORMANCY=(120,500) and HOLD=40. We are running with 6 z/OS lpars in the same JES2 MAS.
Paul Feller AIT Mainframe Technical Support [EMAIL PROTECTED] (319)-355-7824 -----Original Message----- From: The IBM z/VM Operating System [mailto:[EMAIL PROTECTED] On Behalf Of Giovanni Cerquone Sent: Monday, December 18, 2006 1:18 PM To: IBMVM@LISTSERV.UARK.EDU Subject: Parallel Sysplex under VM Hello folks; As part or our DRS test we have implemented two production Parallel Sysplexes under z/VM in a z900-1C7. Parallel Sysplex 1 comprises two virtual CF's and two z/OS guests. Parallel Sysplex 2 comprises two virtual CFs and two z/OS guests. I've been experiencing Start Pendind conditions in both PS. When I look at the online LOGREC I can see lots of MIH entries being registered for the volume where the checkpoint dataset resides. I'm receiving waiting for checkpoint condition avery other time and I've been playing adjusting the Hold and Dormancy values of JES2 with no luck. As matter of fact the MIH entries are being reported for other volumes such as the resident volume and volumes containg coupling dataset. All DASD is sharing and we have GRS Star in place. The Start Pending stuff is killing us and I decided to remove one production system from each PS in order to supress the sharing. So far no more MIH events. Any suggestions as per the MIH value. I found that the MIH value reported by z/OS for DASD is different from the one reported by z/VM. I belive this may be a factor. Kind regards, Giovanni PS. I know that supporting of the Parallel Sysplex under z/VM is not recommended for production but this is what we can do because the machine type we have.
