Nagios Support Forum

I was trying to get the parents directive to work for services but figured out that this bug still exists since April of 2014?
http://tracker.nagios.org/view.php?id=562

Since that appears to be the case, I am trying a service dependency instead. I simply want to ensure that any services that depend on the nsclient++ service to be running don't check or notify if nsclient++ is down:

With this in place, it shows no service dependencies in the UI or when I do a check. However, if I screw up the config, the check fails:
Checking for circular paths...
Checked 19 hosts
Checked 0 service dependencies
Checked 0 host dependencies
Checked 5 timeperiods

And when I screw up the config:
Reading configuration data...
Read main config file okay...
Error: Invalid servicedependency object directive 'shit'.
Error: Could not add object property in file '/etc/nagios/objects/servicedependencies.cfg' on line 2.
Error processing object config files!

Any ideas? Thanks!

What version of Nagios Core are you using?

td3201 wrote:I am trying a service dependency instead. I simply want to ensure that any services that depend on the nsclient++ service to be running don't check or notify if nsclient++ is down:

Code: Select all

    define servicedependency{
            host_name                       .*
            service_description             service-nscp
            dependent_host_name             .*
            dependent_service_description   services-all-auto
            }

This is not a valid service dependency. You cannot use .* and you have not defined any criteria or periods.

Here is an example of a valid service dependency: With this dependency in place, when the service "NSClient++ Version" goes into a HARD w,u,c,p state, the dependent services will not have their checked executed. The next scheduled check time will be pushed back another x as per what is defined in their check_interval. Their service state will remain as is as the checks are not being executed so nothing changes. IF a dependent service was already in a HARD state and was not acknowledged, further notifications will be disabled until the service "NSClient++ Version" goes back into a HARD OK state.

Here is some reading material on service dependencies:
http://nagios.sourceforge.net/docs/3_0/ ... dependency
http://nagios.sourceforge.net/docs/3_0/ ... ncies.html

Box293 wrote:What version of Nagios Core are you using?

Sorry, should have put this in my original post. 4.0.8

As far as getting service dependencies to work, I think I know how to do it. To get around explicitly putting every host in there, I can create a hostgroup and glob in all the hosts, then put this hostgroup directive in the service dependency. I think I've done that before. Will try that for fun.

Service Dependencies is really not what I need here. Just need the parents directive to work. Is that still an outstanding bug?

Here's the working config:

Did this solve your original query?

I'm still curious on the parents directive for services. It's not working as expected on 4.0.8.

I believe that bug may be suspect, and I posted on the tracker for some more information.

Can you share what you are experiencing with service parents?

Also, please post applicable config files and template files.

Thanks.

When the parent service goes down, nagios continues to check and notify about child services:

define service{
use generic-service,24x7,infrastructure,check3
hostgroups windows
service_description service-nscp
check_command check-nrpe-win-service!$HOSTADDRESS$!nscp
servicegroups nscp
}
define service{
use generic-service-notifications-disabled,perf-service,24x7,infrastructure
hostgroups windows
service_description memory pages per sec
parents service-nscp
check_command check-nrpe-win-memory-pages!$HOSTADDRESS$
}
define service{
name generic-service
active_checks_enabled 1
passive_checks_enabled 1
parallelize_check 1
obsess_over_service 1
check_freshness 0
notifications_enabled 1
event_handler_enabled 1
flap_detection_enabled 1
retain_status_information 1
retain_nonstatus_information 1
is_volatile 0
max_check_attempts 3
normal_check_interval 5
retry_check_interval 1
notification_options w,u,c,r
notification_interval 15
register 0
}

I re-confirmed this is still a bug, I'll get a core dev to take a look ASAP.

Thanks

This appears to still be happening in 4.1.0rc1. Any luck identifying the issue?