Pittson, Timothy ,HiServ/US wrote:
>
> Chris,
> This depends on how much trouble you want to go to... the easy way
> is
> to pipe the output from a Q VOL F=D to a file and calculate MB of data
> on each tape based upon percent utilized multiplied by estimated
> capacity, then sort the list.. However, this doesn't account for tapes
> where files span volumes so you'll end up calling for tapes that weren't
> pulled. To account for that takes some additional steps....
>
> 1) Generate the report based upon the number of MB on each tape
> (percent utilized * estimated capacity).
>
> 2) Take all the tapes under a certain threshold (I used 40 MB as the
> cutoff point), and generate ADSM macros to do a Q CONT <volser> COUNT=1
> F=D and a Q CONT COUNT=-1 F=D, piping the output to 2 separate files.
>
> 3) Merge the 2 files by VOLSER, then check for VOLSERs with a first and
> last file segment number equal to '1/1' (otherwise either the first or
> last file spans tapes).
>
> 4) If both the first and last file segment numbers were equal to '1/1',
> the tape would be eligible for processing so write it to a report file
> and also generate the MOVE DATA commands.
>
> I used to go to the trouble of doing Q CONT on the first and last file
> of every ADSM tape, then determining how the volumes were chained
> together but, with the number of tapes we had, this would run for well
> over 4 hours so I gave up on this quickly.
>
> This does sound a bit kludgy (actually it was - took about 16 or 17
> steps if I remember correctly and ran about 20-30 minutes) but it ran
> fine in production for over 2 years prior to migrating ADSM off of MVS
> and it kept our operators happy enought not to carry thru on their
> various threats of bodily harm because of "all of those <fill in the
> your own expletive here> ADSM tape mounts". This was all done using
> some basic utilities (SAS, DFSort, IEBGENER, etc.) so, if you have SAS
> available on your system I'd be more than happy to forward the JCL and
> SAS code to you.
>
> Tim Pittson
> tpittson AT himail.hcc DOT com
> >----------
> >From: Chris Zerr[SMTP:CZERR AT CO.PIERCE.WA DOT US]
> >Sent: Friday, March 14, 1997 11:06 AM
> >To: ADSM-L AT VM.MARIST DOT EDU
> >Subject: Re: Questions on Tape Reclamation -Reply
> >
> >How do you generate the Pull List???
> >
> >>>> "Pittson, Timothy ,HiServ/US" <tpittson AT HIMAIL.HCC DOT COM>
> >03/14/97 06:48am >>>
> >
> >2) Instead of reclaim, use MOVE DATA to move your least full
> >tapes back
> >to disk storagepools - we had a lot of success with this... I used
> >to
> >generate a tape pulllist for the operators of the least full tapes,
> >then
> >would use an ADSM macro to issue the MOVE DATA commands.
> > This worked
> >out well as we'd be able to process 60-90 tapes in an hour using
> >6-8
> >concurrent move processes. You could try doing this one day a
> >week to
> >clear off the least full tapes (i.e. < 25-30 MB used), then use the
> >reclaim method the other 2 days of the week.
> >
> >Tim Pittson
> >tpittson AT himail.hcc DOT com
> >
Me too!
Thanks,
Jane
--
Jane Walker Phone: (801) 237-5532 GO JAZZ! //
Jane Walker Phone: (801) 237-5532 GO JAZZ! //
205 E. 2nd S. Rm 203 Fax : (801) 237-3099 ___//
Salt Lake City, UT 84111 Pager: (801) 480-0175 (_o_/
MailTo: jawalke AT netmail2.mnet.uswest DOT com \\/\//\T//\///
|