Re: 3.4.5 amcheck segfault in try_match with cleaning tape in library

>>>>> "JM" == Jean-Louis Martineau <[email protected]> writes:

JM> That's strange, it means $dev_status == $DEVICE_STATUS_SUCCESS and
JM> $f_type != $Amanda::Header::F_TAPESTART which should not happen for
JM> a cleaning tape.

I've sent a second email with more information, and it seems that if I
undo any changes I made and retry then everything continues to work.

JM> It might be some cache issue, can you send me the
JM> /var/lib/amanda/chg-robot-dev-sg7 file and retry.

I do think you are right.  However, I can't get it back into a state
where it's crashing.  Maybe if I left it in the non-cleaning-tape setup
for an entire cycle, but maybe not.  

JM> At the beginning of volume_is_labelable, can you add:
JM>      debug("volume_is_labelable sl: " . Data::Dumper::Dumper($sl));

I can, but would it still be useful if I can't reproduce the problem?

Sadly I don't have /var being backed up on the backup machine so I can't
get an old copy of the chg-robot-dev-sg7 file.  I realize I should
probably change that.

 - J<