On Thu, Oct 28, 2004 at 10:43:38PM -0500, David Trusty wrote:
> For some reason, there are two sets of files written to the
> same tape, for the same date.  These files are separated
> with (I believe) with some sort of eof marker.

Or, by something else that amrestore *mistakes* for such a
marker.  I haven't looked at the code, but a zero-length file
could possibly have that effect, and your tape does seem to
contain one of those in just the right place :-)

> 3. Do another "amrestore".  This restore can get the 51 disks
> which are in the "retry" dump.

Are you sure the second set contains all 51 DLEs?  It would make
more sense if it contained only the 15 DLEs that weren't in the
first set.  That would be consistent with the log file you
posted, and also with how I'd expect Amanda to work -- if a DLE
failed, Amanda might retry that DLE, but not all the ones it had
already successfully taped!

The output from the *successful* amrestore would settle the
question.  If you don't still have it, you can get the same
output a lot more quickly by running:
    amrestore [opts] /dev/whatever bogus
where "bogus" is a a host-name that does NOT appear on the tape.
That will search for the backups of host "bogus", fail to find
them of course, but print a list of the tape's contents as a side
effect.  Of course you'll have to run that amrestore command
twice, just like the real one...

(Come to think of it, I think there are only 50 DLEs in total on
the tape, since one DLE failed in the estimate phase, and so was
never dumped, let alone taped:
    FAIL planner localhost /mnt/morpheus-home/root 20041012 0 [disk 
/mnt/morpheus-home/root, all estimate failed]
)

> I think I have found a bug in amanda.  I hope someone can
> look into this problem further.  Is there any other way than
> this mailing list to report a bug?

Could well be.  amanda-hackers AT amanda DOT org is a good place.

--

|  | /\
|-_|/  >   Eric Siegerman, Toronto, Ont.        erics AT telepres DOT com
|  |  /
The animal that coils in a circle is the serpent; that's why so
many cults and myths of the serpent exist, because it's hard to
represent the return of the sun by the coiling of a hippopotamus.
        - Umberto Eco, "Foucault's Pendulum"