* Jean-Louis Martineau <martineau AT zmanda DOT com> [20070606 15:14]:
> Try the attached patch, it will gives more debug output on stderr.
> Which changer script are you using? what is its config?

chg-zd-mtx for the glue script. 

Just to make sure, I relabeled all the tape using
a recompiled server and after forcing a full dump I was
getting the same behaviour.

But now I'm getting a debug file and with the patch applied:

# /opt/amanda/spantest/sbin/amfetchdump -p -d /hw/tape/tps6d4nrnsv
spantest wart /bigdisk 20070606 | /usr/freeware/bin/tar -tf -
2 tape(s) needed for restoration
cc
dd (null)
The following tapes are needed: spantest-000001 spantest-000002
Press enter when ready


ee (null)

ff /hw/tape/tps6d4nrnsv
gg /hw/tape/tps6d4nrnsv
Label mismatch, got spantest-000002 and expected spantest-000001
hh /hw/tape/tps6d4nrnsv
Looking for tape spantest-000001...



Looks like amfetchdump doesn't use the fact that the changer has a
barcode reader. One has to manually insert the first tape of the
spanning set in the drive, am I right? If I insert the needed tapes in
the library and leave the tape drive empty it just goes the the 'next'
slot (which btw I don't where it gets it!) and complains of a tape
label mismatch:

hh /hw/tape/tps6d4nrnsv
Could not rewind device '/hw/tape/tps6d4nrnsv': tape_rewind: rewinding
tape: /hw/tape/tps6d4nrnsv: Resource temporarily unavailable
gg /hw/tape/tps6d4nrnsv
Could not rewind device '/hw/tape/tps6d4nrnsv': tape_rewind: rewinding
tape: /hw/tape/tps6d4nrnsv: Resource temporarily unavailable
hh /hw/tape/tps6d4nrnsv
Looking for tape spantest-000001...
ii /hw/tape/tps6d4nrnsv
Label mismatch, got spantest-000007 and expected spantest-000001
hh /hw/tape/tps6d4nrnsv
Looking for tape spantest-000001...
ii /hw/tape/tps6d4nrnsv
Label mismatch, got spantest-000008 and expected spantest-000001
jj /hw/tape/tps6d4nrnsv
Insert tape labeled spantest-000001 in device /hw/tape/tps6d4nrnsv 
and press enter, ^D to finish reading tapes

If I load the first tape needed in the drive everything is good though.
And I assume (not tested) that all the tape needed have to be one
after the others, in the right order. 

jf

> 
> Jean-Louis
> 
> Jean-Francois Malouin wrote:
> >
> >>
> >>The amfetchdump debug file must be in /tmp/amanda/server/spantest,
> >>can you send it?
> >>    
> >
> >no such file:
> >find . /tmp/amanda-spantest -name '*fetch*'
> >returns nothing.
> >  

-- 
<° ><