Re: VTOC size
- From: SKnutson@xxxxxxxxx (Knutson, Sam)
- Date: 10 Feb 2008 08:46:30 -0800
All volumes here are 3390-9 10017 cylinders or the vast majority 3390-27
(LVS) 32760 cylinders. I think the choice to use high capacity tape
more than Large Volume Support impacts the time to recover a single file
though they do go hand in hand often. We want all the economy of scale
we can get when provisioning space and especially when spooling data to
a real media that is in a robot perhaps going offsite. Yes this means we
'waste' 99% of the space on a few 3390-9 volumes that have one of only a
handful of data sets with very strict isolation requirements but the
cost/risk of juggling small volume sizes in our configuration and
recovery configuration is justification. Maybe someday "storage tank"
will be a reality for MVS.
1. Point In Time FlashCopy takes seconds. Backup of FlashCopy target
volume to high capacity 3592 tape much longer but does that matter?
2. Tape once job is complete and the FlashCopy withdrawn
3. Here application or HSM backups are used for more file recoveries
than full volume backups. The Storage team makes it clear full volume
backups taken during disaster recovery syncpoint are for DR not
application convenience. They will recover a file but usually you don't
have someone sitting around waiting for it since the latest full volume
backup tapes go offsite hours after having been written out. HRECOVER
self service is encouraged for users. A little tough love goes a long
way in cutting down the amount of time your enterprise storage team
spends doing individual file restores for folks.
Sam Knutson, GEICO
System z Performance and Availability Management
"Think big, act bold, start simple, grow fast..."
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@xxxxxxxxxxx] On
Behalf Of Skip Robinson
Sent: Friday, February 08, 2008 12:28 PM
Subject: Re: VTOC size
The idea of massive volumes is intriguing, but my concern would be
backup/restore. On some our non-z platforms we have huge volumes many
the size of 3390-3. When one of them went south a while back, it took
and hours to get it restored. With RAID arrays, of course, that's not
supposed to happen; I think it was a problem with an SVC.
On the mainframe we take frequent full volume backups. The likelihood of
true failover (DR) is minuscule, and we use XRC mirroring for that
contingency anyway. On the other hand the probability of having to
an accidentally deleted or modified file is huge--it happens all the
even for sysprog finger checks.
1. How long does it take to make a tape backup of a 32760 cylinder
2. Where does that backup live?
3. How long does it take to restore a critical 10 track file from two
ago that happens to live on the corner of cylinder 31753?
Southern California Edison Company
Electric Dragon Team Paddler
SHARE MVS Program Co-Manager
This email/fax message is for the sole use of the intended
recipient(s) and may contain confidential and privileged information.
Any unauthorized review, use, disclosure or distribution of this
email/fax is prohibited. If you are not the intended recipient, please
destroy all paper and electronic copies of the original message.
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html