On Friday 02 July 2010 13:07:17 Frank Sweetser wrote:
> On 7/2/2010 4:59 AM, Kern Sibbald wrote:
> > You might want to talk to Frank Sweetzer about how he is resolving the
> > problem.  He is also at a University ...
>
> To be honest, I actually haven't been able to fix the crashes yet.  The
> system I'm having them on is quite old - Fedora 8 - so my next step is to
> move up an already planned upgrade to something current, Centos 5.5.  

Yes, CentOS 5.5 is a nice system for us -- it is so widely used that we can 
give pretty good support for it.

> With 
> a little luck I'll be doing the upgrade in the first half of next week, and
> I'll let the list know if it helps my crashes at all.
>
> One slight workaround I did, though, was to reduce my concurrency from four
> to two.  It didn't stop the crashes, but it did reduce the impact a little
> bit - fewer jobs hit by a given crash, and (thanks to spooling) fewer
> volumes with an incorrect file count.

Thanks for the tips -- it is really not a good situation to have Bacula 
crashing.  Possibly after we get Bacula Enterprise 4.0 shipped (early next 
week) we will have time to look at what is going on in zlib.  I suspect that 
they do some system call and don't properly check the return code and handle 
it.  However, in any case they should never seg fault!

Regards,

Kern

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
Bacula-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bacula-devel

Reply via email to