>I have some Very Important Data that I must ensure always exists in at least
>two separate places so that a media failure will never cause a data loss.
...
>I understand that the standard method of doing this is to take a copy of the
>storage pool.  Some of the data might still be on disk, but the amount of log
>data we generate will quickly cause this to migrate to tape.   So we do a
>tape-to-tape copy and have two tapes.  Fine.
>
>Problem:  In the (admittedly rare) case where we're doing the tape-to-tape copy
>and a tape drive decides to eat the input tape, I could end up with ZERO copies
>of one or more logfiles, a Bad Thing.    So I hope to prevent this, somehow.
>(BadThingThreshold=0)

You didn't say what kind of tape technology you're using; but if the data
being backed up is Very Important, then the tape technology being used should
be premium so as to reduce the possibility of tape loss to near zero.  In
several years of backing up to 3590s, I've yet to see the kind of tape failure
you envision.

Another approach is, as another contributor posited, back up first to disk
where caching is turned on, so as to have a copy remain on disk for a time as
it migrates also to tape.  If you're concerned about the disk going bad in the
interim, there is always disk mirroring.

One wonders, also, that if the data is that critical, are the DB2
administrators there mirroring the DB2 log space, to both prevent loss of data
and to essentially eliminate the downtime that would be involved in recovering
the data from *SM?

    Richard Sims, BU