Services of unreachable host

[diverseft]

Hi Guys

My first post here. Hopefully I can explain this clearly

I have recently installed Nagios on a Ubuntu box and everything is working well. I have snmp checks for server hardware as well as generic checks for windows services, mem, cpu etc. All works fine. On top of this i have my "Nagios network" organised with parent/child relationships to tell weather hosts are unreachable to aid in trouble shooting and notification management.

Example:
Firewall - Managed Switch - Server/printer/other device (If firewall is down, managed switch and server/printer are unreachable).

My question is this. I have one site that goes down regularly (ADSL line=poop). The firewall goes down and all hosts behind it fall into a unreachable state so that is good. As all the child hosts are in an unreachable state, what status should the services of the child hosts be?

The reason I ask is that when a site goes down, I only want to be notified about the device that is down. I don't want to be notified of all child hosts services.

What is currently happening is check_SNMP services fall into an unknown state on the printer behind the firewall. services using checks such as 'check_local_procs' or 'check_nt' fall into critical state on the server behind the firewall. In my templates.cfg, I have a template that defines that I am notified of warning,critical and recovery for windows hosts (not unknown), so unknown state is fine as I wont get notified. Critical is where the problem lies. The critical error i get is for example: "CRITICAL - Socket timeout after 10 seconds". I am just wondering weather Nagios should not realise that a parent host is down and therefore put a child hosts services into an unreachable state along with the host it belongs to.

The only thing I can think of is that it is related to that particular plugin that is designed to fall into a critical state if the result of the check is out of range. But I then go back to "should Nagios not check the child host/services until the parent is backup?"

Any advice would be massively appreciated

Thanks

T

[agriffin]

Are you getting notifications about services with down or unreachable hosts? My understanding was that that shouldn't happen. If you are only concerned about the service states, rather than notifications, I can only say that that's expected behavior.

[diverseft]

I am getting service notifications about unreachable hosts.

[agriffin]

Are these active or passive checks? Is the volatile or obsess_over_service options set in their service definitions or templates?

[diverseft]

Hi agriffin

It appears that as the link between the sites drops on and off, the devices are coming on-line again and then unreachable again. I'm guessing its not been detected as flapping as the up down statuses are not changing frequently or quickly enough.....

[agriffin]

I think the volatile option will still have this effect though, even if the services aren't flapping. So is that option set for these services?