Well… since we do not have tape drives available to our IFL, SPXTAPE, every night, is not feasible right now. I can “steal” a drive from z/os when necessary but not every night. It’s a good plan for when we get our VTS (in the works now) and have tape drives available all the time.
We currently have 11 page packs but when I bring up 6.1, there will be 15 or so… We’re not heavy into Linux (yet) so DR has been kind of lackadaisical (sp?) at best… I think I’ll stick to the restore for now, but keep all these great suggestions for later! Anne D. Crabtree System Programmer WV Office of Technology Data Center 1900 Kanawha Blvd East Charleston, WV 25305 (304)558-5914 ext 58292 (304)558-1441 fax From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Mike Walter Sent: Tuesday, August 09, 2011 11:26 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: z/vm page packs at DR NSSes (e.g. CMS, CMSFILES, CMSPIPES, and may more) can easily be backed up using SPXTAPE to write them to tape (preferably every night). Thus the restore at DR is pretty simple (as long as you are very careful to send a copy of the tapes off-site, along with a hardcopy listing of which NSS tape volsers were created on which date). Back when we conducted offsite DR tests, we did just that. The SPOOL volumes at D.R. were created by an EXEC that would autolog a string of userids that had just been added by another EXEC to the DR floor system’s directory. Those autologs permitted concurrent formatting of many DASD, Each autologed ID (conveniently named ‘-volser-‘ where the ‘volser’ was the volume serial number of the volume to be formatted), would format its specific SPOOL, PAGE, or TDISK volume (naturally, just cylinder zero of TDISK vols). A follow-up step instructed how to restore the NSSes from the SPXTAPE backup. A cleanup exec at the end of the D.R. test would carefully delete the ID’s we had added to the floor system’s directory, along with any EXECs and control files that we had added. Mike Walter Aon Corporation The opinions expressed herein are mine alone, not my employer's. From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Hans Rempel Sent: Tuesday, August 09, 2011 9:36 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: z/vm page packs at DR Hi Anne: I have never been in a real DR situation but enough DRs were we did have a time limit to complete. Today we have high capacity/fast tapes so I would follow Clovis suggestion. Start tape restores and put your feet up or ready yourself for the VM IPL once VM base system is restored. One thing you need to consider are the Saved Segments which are in SPOOL space. When the system is shipped most if not all are on the spool section of VM sysres. If any of these get re-saved (via service) they may end up on any SPOOL volume. Therefore make sure all your spool volumes are available before IPL –or be prepared to resave missing segments which is not a good idea in a pressure situation. VM will use SPOOL for paging if not enough PAGE space is available. PAGE is not as critical as SPOOL and can be added later sometime shortly after the IPL (before IPLing of your production systems – since running out of SPOOL space is very bad) using some automated CMS service machine/exec. You need to consider the number of page volumes versus backup time/tape usage. If you have 4 or 5 page volumes I would backup. If you have 20 pages volumes I would consider an exec to reformat them at the DR site. Hans From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of Crabtree, Anne D Sent: August-09-11 9:34 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: z/vm page packs at DR Good point.. Maybe I’ll continue to back them up but for the test, I could IPL as soon as I have one restored (if needed). Anne D. Crabtree System Programmer WV Office of Technology Data Center 1900 Kanawha Blvd East Charleston, WV 25305 (304)558-5914 ext 58292 (304)558-1441 fax From: The IBM z/VM Operating System [mailto:IBMVM@LISTSERV.UARK.EDU] On Behalf Of gclo...@br.ibm.com Sent: Tuesday, August 09, 2011 9:28 AM To: IBMVM@LISTSERV.UARK.EDU Subject: Re: z/vm page packs at DR Anne: I worked with Disaster Recovery for about 10 years. Some observations: "Real" DR are different than "Test" DR: the major factor is Time. On a real disaster you don't have time to think, to plan, to execute... And have a lot of persons making "pressure"... So, the tip is: use the maximum of automation possible. If you have a MVS at the DR site, explore it. From my experience, restore page volumes from tape is faster than format them. You can run several restore Jobs in parallel. To be faster, you can start with a one Page volume and one spool volume, (according to Rick's suggestion: a small spool volume with only the SDF) After the VM IPL, you can add page/spool volumes on fly. Working with tapes, one VM can be live in less than 10 minutes, and you will got time to do the adjusts latter. Good luck, ______________________________________________ Clovis From: Rich Smrcina <r...@velocitysoftware.com> To: IBMVM@LISTSERV.UARK.EDU Date: 09/08/2011 09:13 Subject: Re: z/vm page packs at DR Sent by: The IBM z/VM Operating System <IBMVM@LISTSERV.UARK.EDU> ________________________________ For your initial IPL at DR, you won't need your PAGE packs at all... yes, the IPL will complain. As long as AUTOLOG1/2 isn't kicking off lots of other machines at DR time, you'll be just fine. Then an exec can format the page volumes... re-ipl and all is good. You will need spool space, so you might as well copy it (or keep a small spool area available only with the NSS files, JUST for DR). On 08/09/2011 07:03 AM, Crabtree, Anne D wrote: > > I currently back up all my z/vm packs (res,page,spool) via an adrdssu job on > z/os each > Sunday. At the DR site, I run an adrdssu restore job for these packs. > > This method works fine, however, I’m wondering if backing up the page packs is > necessary? I was thinking that maybe I could backup only 1 page pack so that > I can get > z/vm up and then just init the remaining packs after coming up. Since the > page packs > are listed as cpvols in system config, would z/vm even come up if it couldn’t > find all > of them? Seems like a waste of time to back them up… > > At DR site, we bring up a z/os “rescue” system in order to run restore jobs > for both > z/os and z/vm volumes. Afterwards, our z/vm and z/os systems run as second > level > guests. Maybe I need a “rescue” z/vm system as well? > > Just wondering what everyone else does. > > Anne D. Crabtree > > System Programmer > > WV Office of Technology Data Center > > 1900 Kanawha Blvd East > > Charleston, WV 25305 > > (304)558-5914 ext 58292 > > (304)558-1441 fax > -- Rich Smrcina Velocity Software, Inc. http://www.velocitysoftware.com<http://www.velocitysoftware.com/> Catch the WAVV! http://www.wavv.org<http://www.wavv.org/> WAVV 2012 - April 13-17, 2012 Covington, KY
