I have several NCM Policies that for some time now that show violations but do not turn the policy red for critical alarm as they are setup to do so. What I have noticed is that when I stop the policy and start it again it will then turn the policy red.
Steps to reproduce:
1. Create a policy on a global collection with some devices in it, in such a way that the policy has some violations
2. Repair any existing violation on a policy, so that it turns green.
3. On one of the device on which this policy is already associated, do some configuration changes so that it violates the policy.
Now, we see that even though NCM identifies violator as 1, the policy remains in green color.
The issue is occuring because we are not raising an event on the policy modelhandle when a violation is detected on a device, which is already associated with the policy. When the NCM_POLICY_HAS_VIOLATOR association is created between a device and policy, it should trigger to raise an event 0x82101. Due to this we do not see this event and subsequent alarm on the policy. As a result of which the policy of condition remains normal and color green.
Release : 10.3
Component : Spectrum Core / SpectroSERVER
Fix planned for next 10.3.2 BMP and next 10.4.1 BMP