This problem is being addressed by DB2 APAR IX48568.  I worked with
another customer on this same problem.  The circumvention is to not
use caching in your storage pool and to make sure the storage pool
has a lot of available space.  Just turning caching off in the storage
pool is not enough, you will also have to cause the currently cached
bitfiles to get removed.

There are two methods I can think of to get the cached bitfiles removed.

1 - Set the high migration threshhold to 100 then fill up all of the volumes
    in the storage pool.  This will caused those cached bitfiles to get
    replaced with data.  AFter that you can then set the high migration
    threshhold back down.  (Be ready for tape mounts).

2 - Do a move data for each volume in the storagepool.  After all of the
    data is moved off the volume the cached bitfiles are freed.  You must
    have enough space on the other volumes in the storage pool for this to
    work.

After removing the cached bitfiles (if you were using caching) you still
need to make sure there is a lot of available space in your storage pool.

The way to make sure there is a lot of availalble space in your storage pool
is to set the high migration threshhold relatively low.  Make sure there is
always enough room in the storage pool to hold your largest data base.

Hope this help - David Bohm, ADSM Level 2.