Shad L. Lords wrote:
>> I've been trying 3.0b7 and have found a discrepancy between the
>> documentation and what actually happens.
>>
>> In the documentation for servicedependency it states:
>>
>> dependency_period: This directive is used to specify the short name of the
>> time period during which this dependency is valid. If this directive is
>> not
>> specified, the dependency is considered to be valid during all times.
>>
>> What actually happens is if you don't define a dependency_period is that
>> the
>> servicedependency is never valid.
>
> I may have jumped the gun a little here. What actually appears to be
> happening is when the service is in a soft state checks are sneaking
> through. It may still be related to time but I'm no longer sure of that.
> What is happening though is the following.
>
> Service B depends on Service A.
> Both are on same host and host is up.
> Service B should dependency is defined with execution check of w,u,c,p.
> Both are in Pending status (clean initial start).
> Service A goes to Unknown.
> Before service A goes to a hard state Service B gets checked.
>
> I would think that with Service A's hard state being either U or P (can you
> have a hard P state?) that service B shouldn't be checked. I originally
> though this was related to timeperiods because when I change timeperiods it
> didn't occur. However by the time I restarted after making the change
> Service A had transitioned to a hard U.
>
> -Shad
>
This is actually proper behavior. By default, Nagios will only use hard
states when checking dependencies. This can be changed with Nagios 3 by
setting the 'soft_state_dependencies=1' option in the main config file.
I just noticed this option isn't in the sample config file, and the docs
are outdated, so I'll get that fixed in CVS shortly.
Ethan Galstad
Nagios Developer
___
Email: [email protected]
Web: www.nagios.org
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: [email protected]