<sigh>  This thread, rather than dispelling unsupported speculation, is
now generating it in droves.  It's time to let it go, though I'll
cherry-pick a few points on the way out.


> > The only place where the drive index is stored, AFAIK, is in the
> > FRAGMENT records in the backup's metadata file, as the "dwo" (drive
> > written on) field.
> 
> You've identified it, the "dwo" is the key.

No, it's not.  DWO records where a frag was written and has nothing to
do with allocating a drive for subsequent backups.  That is speculation
on my part, supported by the logic in previous postings and below.  But
I'll bet a bottle of good whiskey on it.  :-)


> Prior to correction the "dwo" assigned from the metadata to 
> the first backup on an 
> assigned tape is now recorded, after the drive number/path 
> correction, the metadata 
> from an assigned tape is now used to mount a tape again to 
> the drive previously used,

I see no reason, in history, logic or experience, to believe NetBackup
(or any other application) works that way, whether you're talking about
drive index, physical drive or index/physical pairing.  If you want to
make the claim that NetBackup does keep mounting the same tape on the
same drive by design, please cite some testable evidence and a
rationale.


> (Metadata (non-dot-f) file information is used for a variety 
> of functions and processes.)

Of course.  But how do you make the leap to "for every tape mount,
NetBackup will search through perhaps half a TB and half a million files
of catalog to find all the frag records involving this tape and pick out
the [presumably identical] last drive index written on [and ignoring
that index/physical-drive changes for a dozen legitimate reasons] and
mount the tape only on that index [even when it's in use, even when it's
down, even when that index isn't in use any more]?"  After more than 40
years of not tieing removable media--card, paper tape, tape, disk,
...--to the particular device it was written on, why on earth would
anyone code that?  Not happening. 

None of us have the source code, but two of you referenced Veritas
advice from support calls.  Would it be too much trouble to make a
follow-up call on your cases for some official info and references?