On Tue, Mar 16, 2004 at 10:58:34AM -0000, Paul Esson wrote:
> I have a NBU v4.5 FP6 set-up on Windows 2000 SP4. At the moment I seem
> to be suffering from a particular issue concerning available media. I
> have noticed that when a job fails having written to a number of tapes
> the media remains assigned to the particular pool that was associated
> with the backup policy and an entry persists in the media db of the
> associated media server. I expected the db cleanup routine that runs
> at the time of a catalog backup to recover this media but it does not
> seem to be happening. I cannot see anything obvious in the problems
> report. Can anyone shed some further light on the matter?

The data that you have written up to the point of failure is perfectly
valid and restorable.  Therefore, NBU keeps it until the images are set
to expire.  There is no real way of it knowing that you are able to
retry the backup, or if that retry was successful.  It may be another
week before that scheduled window is open again.  It plays it safe and
keeps what it's written to tape, and eventually those images will expire
and the media will be available again.

NBU 5.0 has a checkpoint/restart feature that is supposed to allow you
to start where you left off so that will hopefully prevent you from
wasting too much tape.  For example, if you fail in tape 5 of a 10 tape
set, the retry can start half-way through your data instead of from the
beginning.

-- 
Ed Wilts, Mounds View, MN, USA
mailto:ewilts AT ewilts DOT org