You have a couple of options, the quick and dirty solution is to
manually move the data back to the disk stgpool that it came from once
utilization falls to a certain threshold, then allow normal migration
processing to put it back on tape. Depending on how big your disk
stgpool is and what your migration philosophies are, this might work
for you.  A little more sophisticated approach is to create and
designate a disk stgpool for the purpose of reclamation, making sure
that you have a big enough pool to accept all of the reclaimed data
from a single tape, then forcing a migration back to tape after the
tape has been reclaimed.  You would repeat this for every tape that you
wanted to reclaim.  The next step is to automate this process by
writing a script or exec that would query your tape stgpool for
candiate volumes and do the move & migration for you.  There is a UNIX
script named SDRECLAIM in the /adsm/nosuppt directory of the ADSM FTP
site that does this for that platform.  And lastly, there is a
provision in the ADSM V3 server for handling single drive reclamation.
It uses the disk stgpool staging area approach, but what you gain is
not having to write and maintain your own code to manage this process.

Elvin Peterson
epeterso AT scj DOT com