We are in the process of recreating old electronic mail for selected users
in response to a request from our legal department. The mail software we
use makes this a cumbersome process. We restore a couple of top level files
for each user to a system that is normally used for testing mail software
updates. We run a diagnostic utility to get a list of missing files
referenced in the top level files. We restore those files. We rerun the
diagnostic utility to get a further list of missing files referenced in
the restored files. We repeat the diagnostic and restore cycle until we
get a clean diagnostic.

It is fairly easy to reformat the list of files from the diagnostic
utility into a list of restore commands. Unfortunately, the files are
spread over nine tapes and this approach has no particular tendency to
group files by tape. We end up with almost as many tape mounts as there
are files. We can use 'dsmc q backup' to find out when each file was
backed up and sort the requests by backup time. This helps somewhat but
we still get many tape mounts. I can get volume information using the
'show version' and 'show bfo' commands, but 'show version' takes longer
than a typical tape mount (there are over 4.5 million backup files in the
filespace involved).

Is there a faster way to find out which volume a specific file is on?
Alternatively, is there an automated mechanism for converting the list
of files into dsm GUI selections? The mail server and the restores run
under HP-UX. The GUI display usually ends up on a PC running the Exceed
X Terminal emulator.