Unclear on how the "unreachable" logic works

[sbergman27]

Hi,

I have a nagios 3.0.6 configuration that has been working well for me for some time. I decided to define the network topology by setting up 'parents' entries for the hosts. I also removed 'u' from the service and host notification options for the contacts, host templates, and service templates. These are the only places which supply notification options in my config. A couple of hours later, we had an event, and I was still getting notifications for the downstream hosts and services, showing them to be down rather than unreachable.

My original config had max_check_attempts set to 1. I decided that might be the problem and increased it to 2. I have not had an event which would confirm whether or not that helped. But the documentation has me a bit confused. The manual says that setting max_check_attempts should work, but at the expense of performance. So maybe that was not the problem. Also, the option for 'soft_state_dependencies' in the default nagios,cfg I have has it set to 0. I don't understand how the unreachable logic can work with it set to 0. And I'm also wondering if I am confusing 'dependencies' with the 'unreacheable' logic. I've not yet gotten to the part in the documentation which talks about the dependency feature, described as an 'advanced' feature.

Anyway, my goal is to have my notifications to be fairly sensitive to even relatively short outages. But I only want to see the notification for the highest level parent affected, and not for its children, grandchildren, etc.

Any guidance would be much appreciated.

Thanks,
Steve Bergman