To the best of my knowledge, the only way to access inactive
versions if via the "-pick" option, which is what I think you
were referring to when you mentioned "tagging" files. It is
my understanding that the new version of ADSM will include
date filtering options that will help avoid this problem.
We, too, have the same issues as you.

And now a message for IBM: Something that just recently
occurred to me is that we'll need time filtering options
as well, i.e. "-fromtime" and "-totime". This is because
most backups are done at night, and they often cross
midnight. It is my hope that if we don't see time filters
in V2.1, we'll see it soon thereafter in a PTF.

For example, suppose ADSM is running Monday night's
incrementals. This event happens to run past midnight, and in
fact files A.TXT and B.TXT are actually backed up at 00:05,
Tuesday morning.

Next, Tuesday night's incrementals run and A.TXT and B.TXT
are again backed up, except this time they're backed up
at 23:55 (still Tuesday).

Now suppose the user needs to restore A.TXT and B.TXT from
Monday night's incrementals. Using Monday's date would yield
nothing (for the sake of discussion, assume there are no
Sunday scheduled events). Tuesday's date would yield the
wrong versions.

This also implies that we should still continue to use the
-pick option since we can't really know for any given day
whether the backups crossed midnight, or not. My example
whether the backups crossed midnight, or not. My example
of two files is trivial, but when hundreds of files are
involved, the original problems associated with -pick
(although scaled down) still remain.

This differs from "traditional" LAN backups processes where
you have separate tapes for Monday and Tuesday. With separate
tapes, you just pop in the correct tape and the time of the
backup doesn't matter. With ADSM, this is not the case because
we don't separate the media by day.


Andy Raibeck
Connecticut Mutual