Re: Amanda 3.5.1 - "ERROR: amcheck-device terminated with signal 11"

On Sat, Oct 10, 2020 at 23:50:17 -0400, Chris Hoogendyk wrote:
> Wow!
> 
>    amanda@marlin:~/daily$ amcheck -s -otaperscan=taper_traditional daily
>    Amanda Tape Server Host Check
>    -----------------------------
>    NOTE: Holding disk '/amanda3': 913514496 KB disk space available, using 808656896 KB
>    NOTE: Holding disk '/amanda4': 158228480 KB disk space available, using 53370880 KB
>    NOTE: Holding disk '/amanda5': 1636618240 KB disk space available, using 1531760640 KB
>    Searching for label 'Bio-Research-002':label 'Bio-Research-002' not recognized or not found
>    slot 13:slot 13 not in use-slots (14-36)
>    slot 14: volume 'Bio-Research-013' is still active and cannot be overwritten
>    slot 15: volume 'Bio-Research-003'
>    Will write to volume 'Bio-Research-003' in slot 15.
>    NOTE: skipping tape-writable test
>    Server check took 175.512 seconds
>    (brought to you by Amanda 3.5.1)
>    amanda@marlin:~/daily$
> 
> That worked! Interestingly, doing an `amcheck -s daily` after that
> fails just as before. The amanda.conf uses taper_oldest.

Okay, this lends support to the theory that the crash is actually
happening in "scan" operation, rather than in some later part of the
amcheck-driver/taper process.
 
(Were there any error/warning messages written to the amcheck-device log
file for that run?)

> So, maybe if I temporarily go to the different algorithm, it will
> work. Right now the backups are already running and dumping to

Yeah, it might well work, and if so -- and if you don't care which
tape(s) are used next -- then simply switching to taper_traditional
would probably be the easiest approach to getting new dumps actually
written to tape....


If the order the tapes are used does matter to you, I think it should
probably be possible to fix the bug in taper_oldest (oldest.pm) to get
it working (but I'm not really sure how much debugging effort it will
involve...).

If you are interested to attempt that, the next thing I would check is
to see what happens with -otaperscan=taper_lexical (assuming that is also
defined in your amanda.conf).  The "lexical" and "oldest" algorithms
both use the tape-drive inventory (while "traditional" does not), so
that test will help narrow the problem down to just "oldest" or to the
tape-inventory part of the code....  



						Nathan



----------------------------------------------------------------------------
Nathan Stratton Treadway  -  [email protected]  -  Mid-Atlantic region
Ray Ontko & Co.  -  Software consulting services  -   http://www.ontko.com/
 GPG Key: http://www.ontko.com/~nathanst/gpg_key.txt   ID: 1023D/ECFB6239
 Key fingerprint = 6AD8 485E 20B9 5C71 231C  0C32 15F3 ADCD ECFB 6239