Child service becomes reachable, even when only one parent changes to OK

[A41susan]

Describe the bug

As of Icinga version 2.14.0 this change has been implemented: Consider a checkable unreachable once one Dependency fails. Previously all of them had to fail.

This is useful behaviour for our services with more than one parent. But if a service has two or more service dependencies, then it should only return to a Reachable state, when all parent services have returned to status OK.

To Reproduce

Services "parent1" and "parent2" are parent to service "child"

1. Set "parent1" to Critical. "child" becomes unreachable.
2. Set "parent2" to Critical. "child" remains unreachable.
3. Set "parent1" to OK. "child" becomes reachable.

Expected behavior

Since "child" is dependendant on both parent services, it should only be reachable, when both are OK.

1. Set "parent1" to Critical. "child" becomes unreachable.
2. Set "parent2" to Critical. "child" remains unreachable.
3. Set "parent1" to OK. "child" remains unreachable.
4. Set "parent2" to OK. "child" becomes reachable.

Your Environment

Version used (icinga2 --version): r2.14.0-1 Operating System and version: SUSE Linux Enterprise Server 12 SP5 Enabled features (icinga2 feature list): api checker graphite icingadb ido-mysql influxdb2 mainlog notification icingadb --version Icinga DB version: v1.1.0

[slalomsk8er]

As stated in the release note of 2.14.0 under breaking changes:

Redundant dependencies now have to be configured explicitly using the new redundancy_group attribute.

Are any reduncancy_groups involved?

[A41susan]

I'm not using any redundancy groups.

[Al2Klimov]

Hello and thanks for reporting!

What exactly do you mean with becomes reachable? What are the symptoms?

Best, A/K

[A41susan]

Hi,

I'm referring to the indicator on host and service pages in Icingaweb2 under "Check Execution" -> "Reachable" -> yes or no