Thank you for the answers,
splitting is not an option since there will be more like this
I think modifying bacula code would be the go.
As for the devs, they could implement an estimation say 24TB in relation
with 80M/s spooling and writing and so on, and then is way over 6 days.
And then call the off the watchdog.
I am not doing anything wrong :)
Thank you!
------------------------------------------------------------------------
*Mihai Cristian SĂTMĂREAN*
System Administrator
phone:
mobile:
fax:
mail:
web: +49 89 287018 0
+49 89 287018 18
mihai.satmar...@trixter.de <mailto:mihai.satmar...@trixter.de>
http://www.trixter.de
*TRIXTER Film GmbH*
Amalienstrasse 67
80799 Munich, Germany
TRIXTER FILM GMBH Gesellschaft mit beschraenkter Haftung, Sitz Muenchen,
Registergericht Muenchen, HRB 125141
TRIXTER PRODUCTIONS GMBH & CO. KG: Kommanditgesellschaft, Sitz Muenchen,
Registergericht Muenchen, HRA 80561
Persoenlich haftender Gesellschafter: TRIXTER VERWALTUNGS GMBH, Sitz
Muenchen, Registergericht Muenchen, HRB 144097
On 05/28/2013 01:36 PM, Radosław Korzeniewski wrote:
Hello,
2013/5/28 Konstantin Khomoutov <flatw...@users.sourceforge.net
<mailto:flatw...@users.sourceforge.net>>
On Tue, 28 May 2013 10:37:33 +0200
Mihai Sătmărean <mihai.satmar...@trixter.de
<mailto:mihai.satmar...@trixter.de>> wrote:
> lately we try to save on tapes arount 22 TB, and after 6 days and s
> 14-May 18:39 de001bs002-dir JobId 818: Error: Watchdog sending kill
> after 518425 secs to thread stalled reading File daemon. 14-May
18:39
> de001bs002-dir JobId 818: Fatal error: Network error with FD during
> Backup: ERR=Interrupted system call 14-May 18:39 de001bs002-dir
JobId
> 818: Fatal error: No Job status returned from FD. . .
> .
> Elapsed time: 6 days 25 secs
> .
> .
> all jons die exactly after the same time.
>
> Is there a setting to increase the watch dog period, or to make it
> aware that the job is actually running? Can this be a bug?
This is a FAQ question. Bacula has the hardcoded limit of 6 days
for a job to complete. This is not tweakable (and supposedly it won't
ever be):
If someone will write a patch for this it will be tunable. :)
the position of Bacula devs is that if your job takes that
long to complete you're doing something wrong
For "standard" backup jobs it is very true. For some kind of archive
jobs it is not.
-- consider splitting the
task to several jobs.
It is a best known workaround for that. Another is raise a hardcoded
limit and compile Bacula from source.
best regards
--
Radosław Korzeniewski
rados...@korzeniewski.net <mailto:rados...@korzeniewski.net>
------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
