You can update your device class to point to the new library, but
all storage pool using that class will be redirected.

-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]On Behalf Of
Mark D. Rodriguez
Sent: Friday, December 02, 2005 2:12 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: How do you change the device class of a storage pool?


Hi,

I am not sure this can be done, but here is what I would have to do.

    * Customer currently has one library(libA) with one device class
      (devA) defined that points at libA, and several storage pools that
      use devA.
    * Customer has installed new library (libB) which is identical to libA.
    * I would like to change the device class of the copypool to be devB
      so it uses libB and not libA anymore.
    * The upd stg command does not support updating the device class
      attribute.  I can see why it might not, so it keeps people from
      moving an LTO based storage pool to  say a DLT library.

My solution is:

    * rename copypool to oldcopypool
    * create new copypool using new devB class
    * make sure that all new backup data is to diskpool which I backup
      to copypool prior to migration or any data going straight to
      tapepool must goto copypool at the same time via the "copy storage
      pools" attribute.
    * Then start the lengthy backup of tapepool to copypool.  This means
      every byte in tapepool has to be re-copied to the new copypool.
      In my case this will take about 14 days of non-stop spinning of
      tapes on every drive available!  So for all practical purposes it
      will take me 3-4 weeks since I will have to use the drives for
      production purposes during that time.
    * Once tapepool is backed up to copypool I can start the long
      process of deleting oldcopypool.  I will have to do a del vol
      VOL_NAME discarddata=yes for each volume in the oldcopypool.  This
      will be very IO intensive and will take some time as you can only
      run a couple of these at a time without seriously impact the
      performance of the system.

My question:

Does anybody have a better way to do this?  I am missing something in my
process above that might make things simpler?  This is definitely a time
when I wish I could just poke something into the TSM DB!

Thank You in advance for any help you can offer.

--
Regards,
Mark D. Rodriguez
President MDR Consulting, Inc.

===============================================================================
MDR Consulting
The very best in Technical Training and Consulting.
IBM Advanced Business Partner
SAIR Linux and GNU Authorized Center for Education
IBM Certified Advanced Technical Expert, CATE
AIX Support and Performance Tuning, RS6000 SP, TSM/ADSM and Linux
Red Hat Certified Engineer, RHCE
===============================================================================