I made this comment on the RFE. Perhaps it should be a new RFE. For the MOUNT command and in SYS1.PARMLIB(BPXPRMxx), add a NOWAIT and / or DEFER keywords. For a NOWAIT keyword begin the files system check that could take 65 seconds. For a DEFER keyword wait until an access of the file system is requested. In either case, z/OS should process the next command. Any use of the file system would wait until the check is complete. Use on a system file system (list them in the manual (TEMP, ETC, ...)) would be highly discouraged because the IPL will wait until the check is completed anyway.
For a local fix, could another member have a command to process these? Or have automations submit a batch job to run the mount commands? On Thu, Oct 8, 2015 at 11:13 AM, van der Grijn, Bart (B) <[email protected]> wrote: > We have the same issue (more than 2 hours). I've asked IBM about this in the > past and didn't get much of an answer besides WAD. > I've considered taking the application specific mounts out of BPXPRMxx and > mount them after the system is up. Maybe make it part of the application > startup in System Automation. Haven't yet pursued it, but with our DR test > next month it's back on the radar. > The only good thing is that based on the timing of our DR tests this always > gives us a good lunch break. > > Bart > > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On > Behalf Of Dave Butts > Sent: Thursday, October 08, 2015 11:40 > To: [email protected] > Subject: IOEZ00807I when OMVS initializes at Disaster Recovery site > > Searched the archives and saw others had this issue, but never saw a good > answer. > > We recently implemented ZFS file sharing for the first time here. > We run GDPS/XRC to replicate our dasd from HQ to DR site and then often > "flash" the farm to a set we can IPL our recovery system from. > > Because the of replication, our ZFS files are obviously never closed properly > before being used when we IPL the recovery LPAR. > The nature of our 24/7 environment and GDPS/XRC means that the flash of the > dasd farm will ALWAYS be "fuzzy". The contents of block 0 in each ZFS file > will always contain data > > So we experience the dreaded IOEZ00807I message and a 65 second delay per > mount. We have so many ZFS files in use that this takes almost 2 hours to > complete a successful IPL of the recovery system. > > What do other shops do about this? I would hate to have to IPL a minimal > system just to mount/unmount the files properly before IPLing the recovery > PROD system. Is there a parm anywhere to tell ZFS to ignore integrity > checking at IPL? > > Thanks, > Dave > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: INFO IBM-MAIN -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
