On 06/20/2016 12:19 PM, willie bunter wrote: > Hallo to All, > > The dsn was migrated about 2 years ago and for some reason it was deleted > today (in error). I am dusting off some old doc and as I had said earlier I > am checking to see what I did. Keep you all posted if I stumble on to > something. > > -------------------------------------------- > On Mon, 6/20/16, Lizette Koehler <[email protected]> wrote: > > Subject: Re: DFHSM QUESTION - RECOVER DSN FROM ML2 VOLUME > To: [email protected] > Received: Monday, June 20, 2016, 12:41 PM > > If your first statement > is No BACKUP has been taken, then the file is most likely > lost. > > When was the file > migrated? > > Can you show us > the results from a HLIST dsn(/) BOTH > > It will help to know when it was migrated. > > CDS Backups and Journals can > be performed anytime. You just have to know what kind of > performance impact will be felt while it is running > > > Lizette > > > > -----Original > Message----- > > From: IBM Mainframe > Discussion List [mailto:[email protected]] > On > > Behalf Of willie bunter > > Sent: Monday, June 20, 2016 9:35 AM > > To: [email protected] > > Subject: DFHSM QUESTION - RECOVER DSN FROM > ML2 VOLUME > > > > Good > Day To All, > > > > I > would like to recover a dsn which was migrated to ML2. > Because of a mistake > > in the assigning > of the management class no back up was taken. > > > > I remember > recovering a dsn under the same circumstances. I remember > locating > > the dsn in the > HSM.JOURNAL.BACKUP.V000xxxx (found in column 13 in HEX) and > the > > volser of the ML2. (which I found > in column 77) > > > > I > browsed the online HSM.JOURNAL (using the first and second > HLQ) however I > > was unable to locate > the dsn. My question is do I need to wait for the CDS > > Journal backup to be run (which will > execute tonight at 8 p.m) or is there > > > something else I can try? > > > > Thanks. > > > > ---------------------------------------------------------------------- > ... Depending on how desperate you are to recover the data: If you don't delay and (1) you act before the physical tape that contained the ML2 data set is recycled or physically scratched, (2) you save recent backups of the production DFHSM control data sets made from before the ML2 data set was deleted, and (3) you have access to an isolated test z/OS system which has access to tape drives and on which you can bring up a test DFHSM;
then it should be possible to ON THE TEST z/OS SYSTEM: Port the old DFHSM data sets to the test system and bring up DFHSM on those old data sets with everything HELD except for recall; optionally set up an isolated user catalog and alias for the lost data set; catalog (with IDCAMS) the lost data set to VOLSER MIGRAT; set up any ACS definitions and/or storage pools needed to support a recall of the data set (this doesn't need to replicate the production environment as long as the STORCLAS of the data sets maps to a pool with adequate space -- don't know if MGMTCLAS and DATACLAS have to be defined on test system for recall to work); HRECALL the data set to the test storage pool; Use dss to make a tape backup of the data set that can be ported back to production; ON THE PRODUCTION SYSTEM: Read and restore the data set from tape written on test z/OS using dss on the production system; ON THE TEST z/OS SYSTEM: Clean up the no-longer-needed stuff created on the test z/OS system. This is potentially a lot of work and may require some experimentation to get all the pieces right.. I'm pretty sure I actually did this once or twice; but it took the better part of a day and was a long time ago. -- Joel C. Ewing, Bentonville, AR [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
