Well, I just had to do something similar.

What I did is I created a seperate Policy Domain (I am doing absolute (full)
backups once a week of certain systems - I'll explain why below), Policy Set,
Mgmt Class, Archive/Backup copy groups on the server, detailing them with the
data retention values that I wanted.  I then created a seperate node name for
the client (and a seperate Server stanza in the DSM.SYS file on the client) to
work with this Policy Domain.  I also created a different include/exclude file,
to only include the data that I need to backup under this new node name.

Why?  Well, in doing Disaster Recovery Tests, the longest portion of the restore
is the filesystem restores.  With the above configured, I can restore the latest
"full" backup, under the "new" node name, then do the incremental restores from
the "normal" node.  We have had single filesystems take over 18 hours to
restore, since the data is spread over so many tapes in our library.  Colocation
for offsite pools is not feasable in this enviorment.  The customer has the
extra client licences to use, and they never quite understood the "Incremental
forever" paradim of ADSM, so this makes everyone happy (except the ADSM server
and the 3494 library.

Nick Cassimatis
Distributed Systems Management / Storage Management
nickpc AT us.ibm DOT com

If you don't have the time to do it right the first time, where will you find
the time to do it again?