SSD's vs STRASPBAL

I am being told that currently STRASPBAL does not respect the "preferred
storage unit" on CHGPF used to try to keep your file that you want the
fastest access time on SSD.
If you do a STRASPBAL *CAPACITY it can scatter them across SSD's or HDD's.

And, STRASPBAL *USAGE does NOT move "hot" data to less active drives. It
moves dead data to hot drives. Thus freeing up the space on the dead
drives so that new writes can move to the dead drives. My SSD's were at
82% and my HDD's were at 46%. I did a TRCASPBAL and STRASPBAL *USAGE. So
the system said "Gee your HDD's aren't getting a whole lot of writes.
Let's move the dead data over to your SSD's and create 'holes' on your
HDD's. Then any new writes will occur to your HDD's." So now my SSD's
are at 90% with a bunch of dead data.

To fix this I could do a
STRASPBAL *CAPACITY to even out my drives.
Make sure that the PF's are still flagged with a preferred storage unit of
*SSD.
Save the library(ies) off.
Delete them
Restore them.
And then they should be back on the SSD's.

Summary: Do NOT run STRASPBAL, (of ANY kind), if you have a mix of SSD's
and HDD's in the same ASP.


Rob Berendt