On Wednesday 17 August 2005 15:36, Peter Sjöberg wrote:
> On Wed, 2005-08-17 at 14:06 +0200, Jonas Björklund wrote:
> > Hello,
> >
> > On Wed, 17 Aug 2005, viktorija wrote:
> > > tonight i got following situation:
> > >
> > > 17-Aug 01:07 backup-dir: Start Backup JobId 1840,
> > > Job=wise.2005-08-16_22.00.21 17-Aug 01:07 backup-sd:
> > > wise.2005-08-16_22.00.21 Fatal error: Device /baculabk is busy writing
> > > on another Volume. 17-Aug 01:07 wise-fd: wise.2005-08-16_22.00.21
> > > Fatal error: job.c:1665 Bad response to Append Data command. Wanted
> > > 3000 OK data , got 3903 Error append data
> > >
> > > Now i checked all seems ok, but why it happened? What does mean this
> > > error?
> >
> > I guess you running multiple jobs at the same time to different pools
> > on a file device. This is fixed in 1.37.??. Search the list archive.
The problem here depends on what version it is. If it is 1.36.x, then the
problem is that she is trying to "mount" two volumes on the same drive (yes,
Bacula disk backup mimics tape drives ...).
If it is 1.37, the problem is probably fixed in the latest CVS.
>
> Actually, it's probably not yet fixed but is worked on getting fixed in
> a thread on the developers list.
> It seems to be some locking problem and if several jobs that needs to
> change tape starts at the same time one job allocates the device and the
> other fails (when it should wait/share it).
> At the moment my workaround is to make sure the no tape (by executing
> "update slots" which has the side effect of kicking out the tapes from
> the drives) or the right tape is loaded before starting the backups. I
> would stagger the start time but that's to much work for me so I'm
> holding my breath for Kern to be fixed soon (as usual).
I'm working on it and have made a lot of progress. The good news is that I
now have simulated something similar and *finally* know what needs to be
done. The bad news is that some health problems have been dragging me down
for a month now, so programming is going at about half speed. For those of
you who are waiting for responses from me -- sorry, I'll get to them as soon
as I can.
Peter, if you can pull the latest CVS (17Aug05 or later) and run the same test
producing only the SD debug log file as you previously did, it would help.
I've added more debug statements where I was lacking info, and I've,
hopefully, improved the reservation algorithm. I don't expect this version
to fix everything ...
--
Best regards,
Kern
(">
/\
V_V
-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
