More info: Re: [Bacula-users] Getting "Fatal Error: Cannot write block"

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

OK, a little more info on this weird occurrence.

I unmounted the tape in the drive (0010), and mounted the next one (0011).  I 
then initiated a job, and Bacula unloaded 0011 and loaded 0010 again.  But it 
forwarded to the end of the data, and wrote it all just fine.

Do I have a weird one-time glitch?  A failing tape?  Something else?

Here's the report:

05-Oct 14:24 herodotus-dir: Start Backup JobId 6958, 
Job=BackupCatalog.2005-10-05_14.24.24
05-Oct 14:24 herodotus-sd: 3301 Issuing autochanger "loaded drive 0" command.
05-Oct 14:24 herodotus-sd: 3302 Autochanger "loaded drive 0", result is Slot 
11.
05-Oct 14:24 herodotus-sd: 3303 Issuing autochanger "unload slot 11, drive 0" 
command.
05-Oct 14:25 herodotus-sd: 3304 Issuing autochanger "load slot 10, drive 0" 
command.
05-Oct 14:27 herodotus-sd: 3305 Autochanger "load slot 10, drive 0", status is 
OK.
05-Oct 14:27 herodotus-sd: Volume "BaculaTape-0010-20041213" previously 
written, moving to end of data.
05-Oct 14:28 herodotus-sd: Ready to append to end of Volume 
"BaculaTape-0010-20041213" at file=14.
05-Oct 14:28 herodotus-sd: Spooling data ...
05-Oct 14:28 herodotus-sd: Committing spooled data to Volume. Despooling 
196,025,365 bytes ...
05-Oct 14:28 herodotus-sd: Sending spooled attrs to the Director. Despooling 
14,759 bytes ...
05-Oct 14:28 herodotus-dir: Bacula 1.36.1 (26Nov04): 05-Oct-2005 14:28:33
  JobId:                  6958
  Job:                    BackupCatalog.2005-10-05_14.24.24
  Backup Level:           Full
  Client:                 herodotus-fd
  FileSet:                "Catalog" 2005-01-14 05:34:01
  Pool:                   "Default"
  Storage:                "herodotus-sd"
  Start time:             05-Oct-2005 14:24:26
  End time:               05-Oct-2005 14:28:33
  FD Files Written:       53
  SD Files Written:       53
  FD Bytes Written:       195,799,314
  SD Bytes Written:       195,805,757
  Rate:                   792.7 KB/s
  Software Compression:   None
  Volume name(s):         BaculaTape-0010-20041213
  Volume Session Id:      762
  Volume Session Time:    1126137632
  Last Volume Bytes:      13,195,091,792
  Non-fatal FD errors:    0
  SD Errors:              0
  FD termination status:  OK
  SD termination status:  OK
  Termination:            Backup OK

05-Oct 14:28 herodotus-dir: Begin pruning Jobs.
05-Oct 14:28 herodotus-dir: No Jobs found to prune.
05-Oct 14:28 herodotus-dir: Begin pruning Files.
05-Oct 14:28 herodotus-dir: No Files found to prune.
05-Oct 14:28 herodotus-dir: End auto prune.

j----- k-----

On Wednesday 05 October 2005 10:11, Joshua Kugler wrote:
> I sent this yesterday and got no replies.  I'll try to include more detail
> this time.
>
> Server version (SD and Dir): 1.36.1
> Client version(s): 1.36.0 (on all: Linux, Mac, Windows)
> This error is being generated from Windows, Mac, and Linux clients
>
> This error suddenly popped up, with no changes to a working config.
>
> The tape in the drive has a VolStatus of Append, and it only has about 13GB
> written to it (these tapes can do about 50GB).  There are other empty
> (Purged) tapes waiting to be used.  But, I'm getting this error in my
> failing jobs:
>
> 04-Oct 01:05 herodotus-sd: Spooling data ...
> 04-Oct 01:05 herodotus-sd: Committing spooled data to Volume. Despooling
> 64,192,154 bytes ...
> 04-Oct 01:05 herodotus-sd: elive.2005-10-04_01.05.04 Fatal error: Cannot
> write block. Device at EOM.
> 04-Oct 01:05 herodotus-sd: Sending spooled attrs to the Director.
> Despooling 4,400 bytes ...
> 04-Oct 01:05 elive-fd: elive.2005-10-04_01.05.04 Fatal error: Bad status
> 102 returned from Storage Daemon.
> 04-Oct 01:05 herodotus-dir: elive.2005-10-04_01.05.04 Error: Bacula 1.36.1
> (26Nov04): 04-Oct-2005 01:05:08
>
> What is a bad status 102?
>
> If device at EOM, why doesn't it just go on to the next tape?  Do I need to
> manually eject and mount the next one (i.e. is Bacula "stuck")?  Or is
> there something more nefarious that I need to root out here?
>
> But then, even more oddly, the rest of the job report says:
>
>   FD Files Written:       53
>   SD Files Written:       53
>   FD Bytes Written:       184,446,667
>   SD Bytes Written:       184,453,110
> .
> .
> .
>   Last Volume Bytes:      12,999,102,895
>
> How is it writing if it keeps hitting EOM? Last Volume Bytes is not
> increasing across jobs.
>
> BTW, I did a google search for this error, and didn't find anything.  GMAME
> isn't working right now: when I try to search, it says: "No such group:
> gmane.comp.sysutils.backup.bacula.general" even though it lists it here:
> http://news.gmane.org/search.php?match=bacula
>
> I tried to look in the manual for the error codes, but didn't find
> anything.
>
> Any other details I need to post to help out?  Thanks for any tips you can
> give.
>
> j----- k-----

-- 
Joshua Kugler
CDE System Administrator
http://distance.uaf.edu/