Re: Is there something in the system which tells a STRASPBAL on guested disk to bug off?

Rob Berendt <rob@xxxxxxxxx> wrote
on Mon, 28 Nov 2016 14:02:30 GMT:

Found an old Sue Baker reply.
In it she says that STRASPBAL will work with *CAPACITY on
virtual disks but nothing else.
Perhaps that reply was dated and things have changed?
http://archive.midrange.com/midrange-l/200907/msg00827.html
http://archive.midrange.com/midrange-l/200907/msg00921.html

Nothing has changed with respect to virtual disks hosted by IBM
i to IBM i since those posts were made back in 2009 for 7.1 or
later releases. Even with that said, development suggests
opening a PMR and sending the Storage Management VLOGs (major
code 1000) for the time range the TRCASPBAL was running and when
STRASPBAL *USAGE was attempted for debug purposes.

To expand on my previous comment about *USAGE moving cold vs
*HSM moving hot ..... *USAGE uses the trace data to find cold
data to move to highly used units making them more full so new
data goes to lightly used units with the theory being new data
is likely to be hot. Similarly, *HSM will move around both hot
and cold data with a preference to rearranging the data between
HDD and SSD so hot data is on SSD while cold data resides on
HDD.

When all units are of the same type, either should result in
similar resuts and some units will be more full than others.