> I'm planning to do a weekly export of all active files to a tape and keep
> this one offsite in a safe. So in case of a disaster, I should be able to
> restore all the computers back to the state at the beginning of the week. Is
> this a good strategy? (I haven't met any disasters yet and I'm not looking
> forward to it either).

There was a time when this was a very popular approach to offsite disaster
recovery, but most sites abandoned it in favor of copy storage pools when
that facility became available. One of the advantages of copy storage pools is
that they make it relatively cheap to update the offsite information on a
daily basis. A once a week offsite backup runs the risk of losing up to a week
of recent update activity. Few businesses can really tolerate that much
disruption today.

> Beside the offsite tape, what other information is required to recover the
> computer park? I guess that the filesystems of the servers you want to
> restore should match the filesystems before the disaster. Further on, you
> need of course the *sm server and client software. Am I overseeing other
> important information?

Regardless of whether you use export or copy storage pools, you will need a
means of providing each replacement client system with a functioning operating
system and ADSM client software. IBM has a couple of useful red books
available on bare metal recovery. My one major reservation about these books
is that they gloss over the difficulties presented by a replacement system
with a different hardware configuration than the original system (especially
in the case of Windows NT).