We are currently seeing the following situation with the 5.1.5 version of
TSM:

We have multiple backup clients writing to the same primary disk storage
pool.  The migration threshold is reached and a migration is started to the
next LTO tape storage pool.  While the migration process is waiting for a
tape mount; the primary storage pool runs out of space.  The server appears
to randomly select a running backup client to write to the next storage
pool.  The rest of the clients wait for the migration to complete and then
finish writing their backups to the primary storage pool.  At the end of all
of this; we are still left with a backup client writing to the next storage
pool.

Furthermore this client is writing data to the next storage pool at
significantly reduced rate.  The next storage pool has been benchmark at a
write rate of 30GB/hr.  The backup client in this case writes about 41MB/hr.

Should the above be considered normal behavior?  If so is there a way to
improve the write speed?  For example; a parameter that needs to be checked?

Thanks for reply.

Jie