I have my primary storage pools defined as a hierarchy so that backup
files will be initially stored on disk, and then migrated to tape when
the disk storage pool becomes full.

I have the "maxsize" parameter for my disk storage pool set to 16 M and yet
when some client backups run, files much smaller that 16 M are being stored
directly to tape.

For example, yesterday I added a new client and at the end of the backup
session a query occupancy showed:

Type     Filespace       Storage         Number of          Space
         Name            Pool Name           Files       Occupied
                                                             (MB)
----     -----------     -----------     ---------     ----------
Bkup     /               DISK_BU            32,867       1,188.87
Bkup     /               DISK_BU            32,867       1,188.87
Bkup     /               TAPE_BU                 3           8.67

I also did a query content on the tape where the files were written and
it showed:

Type   Filespace   Client's Name for File      Stored Size      Segment   Cached
       Name                                                      Number   Copy?
----   ----------  -------------------------  ------------   ----------   ------
Bkup   /           /usr/local/bin/ dataplot         5.67 M          1/1     No
Bkup   /           /usr/local/bin/ dataplot         5.67 M          1/1     No
Bkup   /           /usr/oracle/dbs/ systora.dbf     2.05 M          1/1     No
Bkup   /           /usr/people/jackal/ AMDbank.ps  991,000          1/1     No

If a client chooses compression, does ADSM use the original file size rather
than the compressed file size to make the decision on which pool to store
the file in ???