[Bacula-users] Bacula don't properly finish it's backups

Hi there,

we are running bacula since several years now. We never had any
problems, which we weren't capable of understanding and solving on
ourselves. But now such a thing happened and I'd really appreciate any
help on this (even ideas!):

We have moved to another ISP with all of our servers (~30). At that
time, we also changed the networking-structure from one big class C net
to several subnets.

All of our servers are running the same bacula-fd version, which are
configured equal, too. All but 2 of them are perfectly working. These
two not working have the problem, that it "seems" as if the backups are
running, but as if they somehow are not finishing correctly.

First, let me show you a mail, we get after each backup from a host,
which is identical in hardware- and network-configuration to one of the
failing nodes:

http://pastebin.com/cdKJ0jua

This is the Mail we get from the failing node:

http://pastebin.com/qzT9tFXw

As you might notice, this trial is running for more than 2 hours and
using several 4 GB - media. So the backup seems to be done. But as you
see in the above quoted mail (which also has the subject "Bacula: Full
Backup Fatal Error fuer emyn-fd") the job is failing.

Here you find the relevant Director - Configuration:
http://pastebin.com/ewGciwqs

This is the SD - Config:
http://pastebin.com/q50Zgd93

This is the (identical, besides the "-fd" names) fd-config of the two nodes:
http://pastebin.com/Aj8sWQQQ

Both fds were configured and build like this:
http://pastebin.com/JXr5w96S

The director's IP is 195.122.136.67
emyn's (failing node) IP is 195.122.135.79
muil's (functional node) IP is 195.122.135.83

Both node's Gateway is 195.122.135.1

There's no firewall or special routing configured for any of the nodes:
http://pastebin.com/aJSQxjrf

The director's version is 2.2.8 .
The fd's version (both/all nodes) is 2.4.4 .

I know, that 5.0.3 has become stable and released, but please
understand, that we can't enroll this version in a live-environment over
30 nodes for testing. Further, it can't be really a versioning problem,
since everything worked before the ISP move. I'm guessing more in the
networking-direction, but have no idea what could explain this or what
to look for.

I'm just running out of ideas here.

Thanks in advance for your help!

Best regards,
Marc


------------------------------------------------------------------------------
Lotusphere 2011
Register now for Lotusphere 2011 and learn how
to connect the dots, take your collaborative environment
to the next level, and enter the era of Social Business.
http://p.sf.net/sfu/lotusphere-d2d
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users