-
Bug
-
Resolution: Done
-
Normal
-
netobserv-1.6.1
-
None
-
False
-
None
-
False
-
-
-
NetObserv - Sprint 259, NetObserv - Sprint 260
-
Moderate
-
Customer Facing, Customer Reported
OCP 4.16.0
With Network Observability up and running, do the following:
oc label namespace netobserv test="true" oc label namespace netobserv openshift.io/cluster-monitoring="false" --overwrite
Enter "oc get namespace netobserv -o yaml" to view the labels on the netobserv namespace, or in OCP Web console, go to Administration > Namespaces and select "netobserv".
Now delete the netobserv-controller-manager-* pod in the openshift-netobserv-operator namespace. It will restart. After about 20 seconds, the label is reverted back to openshift.io/cluster-monitoring="true" and the custom label test="true" is removed.
—
This issue only happens when you change openshift.io/cluster-monitoring="false", or something occurs that triggers the reconciler code to run and reset the labels.
---
Suggested solution:
In the next release, netobserv will add controller ownerReference to the created namespace YAML and will reconcile as soon as this reference match.
To override the YAML, simply remove the owner reference field and do your update.
This solution apply for any object owned by the controller.