Bacula-users

Re: [Bacula-users] lots of copy jobs fail suddenly (5.2.12)

2012-12-17 08:49:28
Subject: Re: [Bacula-users] lots of copy jobs fail suddenly (5.2.12)
From: Uwe Schuerkamp <uwe.schuerkamp AT nionex DOT net>
To: Bacula Users Mailing List <bacula-users AT lists.sourceforge DOT net>
Date: Mon, 17 Dec 2012 14:47:19 +0100
On Thu, Dec 06, 2012 at 03:47:17PM +0100, Uwe Schuerkamp wrote:
> On Fri, Nov 30, 2012 at 11:56:23AM +0100, Uwe Schuerkamp wrote:
> > Hi folks,
> > 
> > after a few rather uneventful weeks, all of a sudden some copy jobs
> > (no pattern discernible) have started to fail with messages like
> > these: 
> > 
> > 30-Nov 09:55 deniol186-dir JobId 51675: Copying using JobId=51200
> > Job=deniol2147.2012-11-27_22.05.57_09
> > 30-Nov 09:55 deniol186-dir JobId 51675: Storage "" not found, using
> > Storage "FileStorage_server2147_full" from MediaType "File_server2147_full".
> > 
> 
> Hi folks,
> 
> the above issue still persists, random copy jobs fail because bacula
> cannot correctly determine which read storage to use.
> 
> the jobs finish successfully if I run a job manually and set all the relevant
> params (pool, storage) in the bconsole dialog. 
> 
> Is there a way to enter a list of jobids to copy in bconsole by using
> some special copy job selection pattern definition? Right now I have
> to edit bacula-dir.conf and reload the new config, using a single
> job_id to copy in a "static" sql select statement 8-P 
> 

It looks like the error has to do with some recent "optimization" I
added to the select statement: I had wanted to sort copy jobs by job
size so the big jobs would run first (during the night) so it would be
easier to stop the copy jobs during the day if access was needed to
the lto5 drive. 

Still we have many copy jobs failing, but not as many as before as we
now randomize job order more or less. Still I think it's a bit
brain-dead by bacula that it's unable to sort out which storage and
pool a certain job used during the online backup, so all jobs wait on
the general "File" storage instead of using the correct storage
device. 

I've heard it on the grapevine that some major changes to copy job
logic and / or disk-based volume mgmt might trickle down into the
community version, any more info on this? 

Is there a way I can manually trigger a copy job via text file where I
can provide the job id, storage and read pool to be used? 

All the best, 

Uwe 

-- 
NIONEX --- Ein Unternehmen der Bertelsmann SE & Co. KGaA



------------------------------------------------------------------------------
LogMeIn Rescue: Anywhere, Anytime Remote support for IT. Free Trial
Remotely access PCs and mobile devices and provide instant support
Improve your efficiency, and focus on delivering more value-add services
Discover what IT Professionals Know. Rescue delivers
http://p.sf.net/sfu/logmein_12329d2d
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users