In response to some recent postings:

Over the years, I have built a series of Perl scripts on my AIX
system to control the operation of TSM.  The scripts have reached
a high level of sophistication in what they can accomplish.  If
anyone wants these, just e-mail me and I will send them out.

Basic structure is:

At appropriate times during the day, cron processes run scripts to
load a file with a list of tasks to be performed.  Each entry includes
a list of requirements that must be matched in order to run the
task.   These requirements include such things as number of tape
drives required, do not run if storage pool xxxx is migrating, do not
run if storage pool yyyyyy is being used by any task, do not run if
a database backup is running.

Every two minutes, a cron process runs the main script which
issues a series of query commands into TSM and parses the
answers to determine
  - what tasks are running
  - how many tape drives are defined and online
  - how many tape drives are in use
The script then looks through the list of tasks to be performed and
checks each one to see if it can be run based on its list of
requirements.  The first entry that matches is run.  The script then
exits.

The task that gets run is actually a command issued to AIX.  The
command is usually a script that calls TSM with the desired
command.

The scripts that build the list of tasks do different things based on
the day of week, and if the day is a holiday.

There is a script that checks all the tape pools and determines
which pools have the most tapes that need reclamation.  The list of
tasks is then built with the reclamation commands ordered to
process the largest first.

Warning:  The scripts have been built based on our naming
conventions for tape drives and storage pools.  Adapting the main
script to your environment will require some reprogramming.  The
good news is that the scripts contain lots of comments to explain
what is being done at each point.

Since the scripts have been built over several years, some of the
programming could be made fancier.  For example, use a SELECT
command for the desired fields rather than a QUERY
FORMAT=DETAILED and parse through all the text.

So, if you are interested just e-mail me at     mhull AT uottawa DOT ca
and I will e-mail you a tar file.  The file is just over 100K in size.




Mike Hull               mhull AT uottawa DOT ca
Systems and Automation Analyst      613-562-5800 x 4587
Computer Centre        /  136 Jean-Jacques Lussier Room 543
University of Ottawa   /  Ottawa, Ontario   /  Canada   K1N 6N5