log meesage antrea-mc-controller

[AdeshGhadage]

This PR addresses issue #6152

I am unable to verify if the issue is resolved. I have modified the order and placed the cleanUpExpiredMemberClusterAnnounces function to delete any MemberClusterAnnounce if it's after // SetupWithManager sets up the controller with the Manager. However, the following error still appears:

E1213 18:47:35.881379       1 stale_controller.go:71] "Failed to get MemberClusterAnnounces" err="the cache is not started, cannot read objects"

It should instead show:

E1213 18:47:35.881379       1 stale_controller.go:136] "Failed to get MemberClusterAnnounces" err="the cache is not started, cannot read objects"

To recreate and verify the log, I followed these steps:

make clean
make antctl
minikube delete -p leader-cluster
minikube start -p leader-cluster
kubectl config use-context leader-cluster
kubectl create ns antrea-multicluster
./bin/antctl mc deploy leadercluster -n antrea-multicluster --antrea-version $TAG
./bin/antctl mc deploy membercluster -n kube-system --antrea-version $TAG
kubectl get all -A -l="component=antrea-mc-controller"
kubectl logs -n antrea-multicluster deployment.apps/antrea-mc-controller -c antrea-mc-controller

[luolanzone]

May I know why do you create a duplicate PR the same as before? #6645

[AdeshGhadage]

May I know why do you create a duplicate PR the same as before? #6645

Apologies for the confusion! I created this PR because I was unable to verify the fix in the previous PR (#6645), and the issue still persists. The author of the earlier PR mentioned they were not able to recreate the error locally, but I have successfully recreated it on my setup. I also modified the order of cleanUpExpiredMemberClusterAnnounces, but the log still shows the same error. I opened this PR to highlight the issue and seek further feedback or clarification. If you'd prefer I continue the discussion on the original PR, I can close this one.

[luolanzone]

OK, No problem, we can keep this to start a new discussion.