Bacula-users

Re: [Bacula-users] SD crashes

2012-02-13 09:28:23
Subject: Re: [Bacula-users] SD crashes
From: Joe Nyland <joenyland AT me DOT com>
To: bacula-users AT lists.sourceforge DOT net
Date: Mon, 13 Feb 2012 14:24:19 +0000 (GMT)
On 13 Feb, 2012,at 11:37 AM, Adrian Reyer <bacula-lists AT lihas DOT de> wrote:

Hi Joe,

On Mon, Feb 13, 2012 at 07:21:03AM +0000, Joe Nyland wrote:
> I hope someone would be able to offer any suggestions of why I am seeing the following behaviour in my current Bacula setup:
> Since the tail end of last week, I have been having issues with my MySQL backups in Bacula, where they would randomly appear to 'crash', normally when performing a copy of a backup to another pool - but I'm not sure yet if this is the trigger.

With bacula 5.0.3 I had frequent crashes on Copy Jobs as I ran out of
memory. The SD-box has only 4GB RAM, now I added 8GB swap and it seems
to run fine.

> NOTE: bconsole appears to crash here - no further output is produced, and bconsole does not respond to any key presses. I have to Ctrl + C to exit out from bconsole. Furthermore, the only way I can clear our the failed jobs from the 'Running jobs queue' is to exit from bconsole, issue 'sudo service bacula-sd stop' twice, then restart the SD and restart bacula-director.

Here the bacula-sd crashes and misses from process list.

I have another issue I have not been able to track down so far. The tape
changer seems to claim it has 0 slots now and then and bacula-sd really
dislikes that. Seems mostly to happen when tapes are moving and some
'mtx status'-like command is issued. If this happens, I need to stop
bacula-sd, it will take some time to umount the tape (bacula-sd has 'D'
state in 'ps'), only afterwards it can be started again and all is fine.
'update slots' without restart won't help, even as 'mtx status' gives
correct output again. Perhaps this is comparable to your "issue 'sudo
service bacula-sd stop' twice".

Regards,
Adrian
--
LiHAS - Adrian Reyer - Hessenwiesenstraße 10 - D-70565 Stuttgart
Fon: +49 (7 11) 78 28 50 90 - Fax: +49 (7 11) 78 28 50 91
Mail: lihas AT lihas DOT de - Web: http://lihas.de
Linux, Netzwerke, Consulting & Support - USt-ID: DE 227 816 626 Stuttgart
 
Hi Adrian,

Thanks for your reply.

I hadn't considered RAM as being the cause of the problem, mainly because other backup jobs backup far more (and far larger) files to this same SD without issue. It seems to be only when I introduced MySQL backups of different servers (including the Bacula catalog server) into the mix, that I started to see this behaviour.

My current theory which I am testing is disabling the MySQL backup and copy jobs for FileServer1 only, so that the Bacula database is not backed up in Bacula as this resides on FileServer1 - I'm starting to wonder whether the process of backing up my catalog at the same time that several other backup jobs are running (and completing in the case of the smaller DBs) is somehow causing this problem. However, this doesn't explain why the SD appears to be crashing :-(.

In the meantime, I have found this bug which was forwarded on from Debian bugs: http://bugs.bacula.org/view.php?id=1098, However it appears to for Bacula 2.2.8 :-( Another mention of the issue here: http://adsm.org/lists/html/Bacula-users/2009-12/msg00140.html but that's for Bacula 3.0.3.

Any other ideas?

Thank you.

Joe






------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users