Re: [Shinken-devel] Services Depencies question

[nap <nap...@gm...>]

Hi,

Sorry for being long, the 0.8 release took me a lot of time to get done.

On Thu, Oct 6, 2011 at 6:23 PM, François Picot <ml...@pi...> wrote:

> Hi!
>
> I'm still playing with depencies and escalations, without luck.
> A new bug I found is that "notification_failure_criteria" doesn't seems
> to have any effect.
>
> With 2 dummy tests, and this dependency :
> define servicedependency{
>         host_name                       monit
>         service_description             test1
>         dependent_service_description   test2
>         execution_failure_criteria      o
>         notification_failure_criteria   o
> }
>
> If both test1 and test2 are critical, ant test1 goes OK, test2 is
> checked one last time (it should'nt but it's not too bad), but
> notifications for test2 keep getting sent. And if test2 goes OK, as it
> isn't checked anymore, it isn't detected, ant notifications are still sent.
>

Oh, It seems that the "OK" state is not took in the dependency logic I
think. I'll wrote a test case for it, but it sounds like a bug because when
we wrote the dependency logic, it was in ming the "bad cases", not the good
ones.


> But as you asked in bug #317, I'll try the latest git version before
> anything. I'm still in 0.6.5, and that might explain a lot.
>
> Le 06/10/2011 16:22, nap a écrit :
> > Hi,
> >
> > What if :
> > * you remove the execution_failure_criteria line?
> > * define ServiceA in the service_dependencies property of the service A?
> I'll try that if upgrading isn't enough.
>
> > >   Will serviceB be dependent upon serviceA on the same host, or will it
> be
> > >   dependent upon serviceA on any host in hostgroup1.
> >
> > It will be the first one :)
> OK, that's great.
>
> Next question is : how about a service which should run at least on one
> host? would the following depency do the trick?
>
> define servicedependency{
>         hostgroup_name                 group1
>        service_description             test
>        dependent_hostgroup_name        group1
>         dependent_service_description   test
>         execution_failure_criteria      o
>         notification_failure_criteria   o
> }
>
No I don't think so. The group will be just like N dependency in the end,
not N*N. Maybe a "business rules" can do the trick with an "OR" rule?

Regards,


Jean


>
> >
> > Regards,
> >
> >
> > Jean
>
> Best regards
> François
>
>
> ------------------------------------------------------------------------------
> All the data continuously generated in your IT infrastructure contains a
> definitive record of customers, application performance, security
> threats, fraudulent activity and more. Splunk takes this data and makes
> sense of it. Business sense. IT sense. Common sense.
> http://p.sf.net/sfu/splunk-d2dcopy1
> _______________________________________________
> Shinken-devel mailing list
> Shi...@li...
> https://lists.sourceforge.net/lists/listinfo/shinken-devel
>