=> On Wed, 24 Sep 2003 10:21:05 -0400, Bill Fitzgerald <wfitzger AT MHC DOT 
NET> said:


> when I am running the backup of my primary disc storage pool to tape, I use a
> max process of 4. I make sure that there are plenty of scratch tapes available
> and that the max scratch value for the copy pool is larger than the used. so
> tapes are available. the copy pool is non collocated.

> What I am trying to find out is why
> 4 processes start correctly and run. over time, the jobs end successfully so
> I end up with only one drive backing up data. The jobs that complete do so
> when there is still data to be copied to tape.

> Why would they end if there are tapes available and data to be copied?


Each process only manages one node at a time.  So, so long as there are at
least four nodes to be moved, you will maintain four processes.

It is my anticipation that you have one node which is significantly larger
than most of the others.  That's the one that is last to finish.

What you can to to evaluate this is run a 'q occ stg=YOUR-DISK-POOL'
periodically as you go.  You'll find many nodes represented, but then when you
see one of the processes end, it'll go down to 3.  Then 2, and then 1.


- Allen S. Rout
- UF TSM type