Mauro,
that's quite a philosophic question. And you have two possibilities:
Request a backup to be completed BEFORE HSM does the migration. Then
there's no need to backup the subfile again. For ADSM HSM is transparent, so
ADSM doesn't even get an idea that the file in question could already have
been migrated. Thru incremental forever method you don't have to read the
migrated file again (as it doesn't change).
Make the stubfiles visible to ADSM. With this method you'll have the
original file in your backup until it gets migrated, then it gets replaced
by a stubfile. In case of a server failure you only have to recover the
stubfiles, that's fine. When your HSM environment on the server crashes
you're quite dead, that's not fine... 

I think both methods have their advantages. When one of my servers crashes
and has to get replaced, the new one will be larger than the old one.
Probably all HSM data will fit on the new machine.

Best regards,
Michael
--
Michael H. Bartl
Michael H. Bartl
ADAC e.V. (RZ/PPO, Storagemanagement, Datensicherung u. Notfallvorsorge)
Tel.: (089) 7676-4063, Fax: (089) 7676-8161
mailto:michael.bartl AT zentrale.adac DOT de
http://www.adac.de