Bump. Nobody has had this issue before? :-)
For what it's worth, I have "hard status only" enabled for the
dependencies, and the services are configured for only 1 check for hard
status.
On 4/21/2015 1:08 PM, Russell Jones wrote:
Hi all,
I have configured the following service dependencies:
define servicedependency{
dependent_hostgroup_name rdhpc-nodes
dependent_service_description NX NXD Service Check,NX NXNode Service
Check,NX NXServer Service Check,SSH Check,NIS Client Service
Check,Munge Service Check,Slurmd Service Check,LSF Service Check
host_name evxcat
service_description TFTP Service Check
execution_failure_criteria w,u,c
notification_failure_criteria w,u,c
}
I am attempting to test this to ensure it works as expected. I am
stopping the TFTP service on my evxcat host and the LSF service on one
of my rdhpc-nodes hosts. Unless the "TFTP Service Check" runs before
the "LSF Service Check" I will get an alert for the LSF service even
though the TFTP service on the evxcat host is down. It is my
understanding that "enable_predictive_service_dependency_checks=1" is
supposed to have Icinga launch a parallel check of the TFTP service
check to ensure it is in a hard "OK" state before sending out a
notification for the "LSF Service Check". This is not happening.
cached_service_check_horizon is set to 15, and the checks are not
happening until several minutes later so that TFTP service "OK" status
should not be cached.
Any assistance in fixing this is greatly appreciated!
_______________________________________________
icinga-users mailing list
[email protected]
https://lists.icinga.org/mailman/listinfo/icinga-users
_______________________________________________
icinga-users mailing list
[email protected]
https://lists.icinga.org/mailman/listinfo/icinga-users
