Bacula-users

[Bacula-users] Connection to Storage daemon hangs

2009-09-03 11:08:44
Subject: [Bacula-users] Connection to Storage daemon hangs
From: Markus Kress <Markus.Kress AT gls-itservices DOT com>
To: bacula-users AT lists.sourceforge DOT net
Date: Thu, 3 Sep 2009 16:39:49 +0200
Hello,

I have a problem with Bacula 3.0.2 running on Linux. All bacula daemons
running on the same system.

The problem is occurs while doing a "status storage". The connection to the
SD seems to hang.

Connecting to Storage daemon tape at degpn015w156:9103
<... nothing more happends for ever>

Restarting the DIR solves the problem. Restarting SD does not help. So I
think the problem is not related to the SD.

How the problem occur:

I have a admin job right before backup job (scheduled at the same time but
with different priorities), which do some tests. In some cases it unmounts
the tape (LTO2) with the bacula "unmount"  and eject the tape with "mt -d
/dev/nst0 offline". The following backups will fail, because no tape is
mounted. That's how it should work.

The first Backup after unmount/eject  fails sometimes which this message:

18-Sep 02:18 degpn015w156-director JobId 110: Fatal error:
Storage daemon didn't accept Device "tapedrive" because:
3924 Device "tapedrive" not in SD Device resources.
18-Sep 02:18 degpn015w156-director JobId 110: Fatal error: Max wait time
exceeded. Job canceled.
18-Sep 02:18 degpn015w156-director JobId 110: Error: Bacula
degpn015w156-director 3.0.2 (18Jul09): 18-Sep-2009 02:18:28

"Max wait time exceeded" is what I expected after eject the tape, but what
about the message "3924 Device "tapedrive" not in SD Device resources?

"tapedrive" is in the config!

The admin Job (which sometimes unmount/eject a tape) and Backup-Job starts
at the same time with different priorities. If I insert in my admin job
some sleeps (after eject for instance), the problem still exists. But if I
start them manually job after job, I never got any problems. Is there a
internal lock problem of bacula?

My workaround will be changing the scheduled start ime and disabling the
backup jobs, by changing the job config dynamicly (to "Enabled = No" and
reloading the config). After some hours I reenable the backup jobs again.
But I'm still interessted for a other solution (bug fix?).

Mit freundlichen Grüßen / Best regards

Markus Kress
System Technologies

GLS IT-Services GmbH
GLS Germany-Straße 1-7
36286 Neuenstein
Germany

T +49 (0) 66 77 17 426
M +49 (0) 172 1781 426
F +49 (0) 66 77 17 486

E markus.kress AT gls-itservices DOT com


___________________________________________________________________________________________________________________

GLS IT Services GmbH
Sitz: Neuenstein, Amtsgericht Bad Hersfeld HRB 388, Geschäftsführer: Rüdiger 
Schmahl
___________________________________________________________________________________________________________________

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users

<Prev in Thread] Current Thread [Next in Thread>
  • [Bacula-users] Connection to Storage daemon hangs, Markus Kress <=