On the very rare occasion (like once in 15 years) where we really needed
to recover data off a ML2 tape using backup CDS data sets, We ran our
stand-alone "one-drive" test system in a SysProg test LPAR to restore
the related CDS datasets, fire up DFHSM, recall the needed datasets off
of the ML2 tape to test system volumes, and port the datasets by some
means back to the system that needed them (where they could be
re-migrated if appropriate). That way you avoid running DFHSM with old
CDS data sets that are inconsistent with a production environment, with
access to production datasets , and all the considerable risks that entails.
In our case it wasn't a hardware problem, but by some accident an
important dataset on ML2 got deleted, yet we were able to locate a
physical ML2 tape still containing the logically-deleted dataset and
corresponding backup copies of the CDS datasets that reflected its
presence on the tape. In those days our tape drives weren't in a
library, so varying a single drive on-line to a stand-alone test LPAR
was easy.
If you can identify the ML2 data sets that are involved, the first step
is is ask the data set owners whether they can tolerate the loss of the
data set or whether they have easy means of rebuilding the data. Some
datasets are retained just-in-case and may have already passed their
potential usefulness. We ran into several cases over the years where
Applications Programming decided they had better alternatives when they
learned recovery of a specific "lost" data set could potentially be done
from our DR volume backups or by DFHSM "time travel", but would require
some major magic on our test LPAR and the better part of a work day.
Joel C. Ewing
On 7/12/22 21:39, fred glenlake wrote:
Hello Listers,
I have a small remote site that we are upgrading the physical tape library to a
virtual TS7700 library. All went well until the old library had a serious
hardware error. My leadership has decided not to spend the money to repair
the old failed library (of course now out of warranty and maintenance).
However, I still have one more ML2 tape worth of data to migrate to the new
library.
If we bring the ML2 tape back to our main site would it be possible to read the ML2 tape to extract
the data off of it? The only way I can think of is to also bring copies of all the CDS's and
Journal from the remote site, restore those CDS's (MCDS, BCDS, OCDS) and Journal. Then bring down
the "real" DFHSM and bring up the "Remote" DFHSM to be able to recall the data
off of the ML2 tape.
Unfortunately running the "hsend list dsname both" commands on the ML2 datasets
shows No BCDS records so I cannot just recover the ML2 datasets from BCDS copies.
Any suggestions would be welcomed.
Thanks,
Fred G.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
--
Joel C. Ewing
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
