Don,
You do have a valid concern regarding duplicate volsers after FLASHCOPYing
entire dasd volumes. However, its very easy to clip the volume to an
alternative volser (almost) immediately after the FLASH (CPFMTXA vdev
nvolsr LABEL). Our Disaster Recovery processing does just that. You're
theoretically exposed to duplicate volsers for a few seconds (the amount of
time we wait to allow the FLASH to occur) but we've never run into a
problem with that. It would be nice if FLASHCOPY did provide the option to
relabel but it doesn't. That would make a good SHARE requirement.
I believe the allocation maps are stored on cylinder 0 so you'd be losing
that data by FLASHing starting at cylinder 1. So, I'd just go ahead and
start w/ cyl 0.
If you've FLASHd the entire volume to a backup, you could theoretically
flash the backup back to the source (with a relabel, of course). But you
need to concern yourself with synchronization of data with other volumes
(can I restore one volume safely without restoring other volumes?). So,
for example, if you were restoring a sysres volume, you'd need to clip the
label back the original sysres volser and be aware of the implications of
unsynchronized data (spool, for example). Thats the same issue you'd have,
though, if you were just dumping to tape instead of FLASHing; FLASHing
makes it less of an issue.
For our DR environment, as an example, we FLASH all of our volumes (w/
relabel), backup the relabeled volumes to tape, then restore the tape data
to volumes labeled with their original volsers. Then IPL.
Let us know if you have any further questions.
Dennis Schaffer
Mutual of Omaha
"Don Abbate"
<[EMAIL PROTECTED]
tate.va.us> To
Sent by: "VM/ESA [email protected]
and z/VM cc
Discussions"
<[EMAIL PROTECTED] Subject
.UARK.EDU> VM system backups
01/29/2006 05:38
PM
Please respond to
"VM/ESA and z/VM
Discussions"
<[EMAIL PROTECTED]
.UARK.EDU>
We have always backed up VM system packs to tape using DDR. With the
advent of Shark and flashcopies we are thinking about using flashcopy
instead. But that brings up the problem of duplicate volsers if we do the
entire pack. We are thinking of labeling the target volumes something
different from the system volsers and then flashing from 1 to end instead
of 0 to end. This will get around the duplicate volser, but do we lose
anything by not getting cylinder 0? Where are the pack allocation tables
kept? What problems might we run into by flashing back to the original
volume or if we have to flash to a totally different volume? Is it
possible to clip the target volume to the system volume names and just IPL
off that if need be? Any other recomendations?
It would certainly be nice for the flashcopy command to have another
parameter that could specify a new volser for the target volume so we could
flash 0 to end and not have the duplicate problem.
Thanks.
Don
