Rick got it right.  In our library, after the database backup is made and
back in its slot, we issue:

        move drmedia * wherest=mountable tost=courier rem=bulk

That pulls the database tape and copypool tapes that are in the library,
checks the volume name on each volume and puts it in our output door (our
library is a 3583).
Once it is done, all the tapes to go offsite are listed in the stat
'courier'.
We do a

        q drmedia

Before tape cases come in from offsite (we do it a case at a time, and
before we got a little to big, one case per day with 20 tapes it it was
enough), and pull any tapes listed as 'vault retrieve' as well as the
expired database backup tape (we have both the reuse delay and db backup
expiration set to 6 days).

For each tape that was found, we do a

        move drmedia VOLNAME wherest=vaultr

This moves the tapes specified to the status of 'courier retrieve'.
Then we issue both:

        move drmedia * wherest=courier
        move drmedia * wherest=courierr

Now the database tape and the retrieved tapes can all be checked in as
scratch.

Understanding the state machine used in DRM is not hard, but it does take a
little study to be comforatable with it!