This message was deleted.

Start here:

kubectl --namespace kube-system describe deploy/coredns

for me it usually works to execute this command to check the status of the kube-system pods;

kubectl get all -n kube-system

And if it turns out that one or more pods are being a pain in the *, then I'm pretty much forcing them to redeploying.

kubectl delete --all pods -n kube-system --force

@kind-nightfall-56861 that works sometimes. But instead of deleting pods, you could restart the deployment that controls the pods:

kubectl rollout restart -n kube-system deployment coredns

But you might want to find out why the deployment is failing before you do that. That's what I suggest:

kubectl --namespace kube-system describe deploy/coredns

Tbh, when I try do restart through the Rancher interface, restarting almost never works, but idk if that restart translates to that console line. I'm finding that my method works 100% of the time, but I might be mistaken.

Deleting the pod is almost like restarting a deployment that has

.spec.strategy.type==Recreate

. I think the default is

.spec.strategy.type==RollingUpdate

. The latter will try to start new pod before stopping the old one, which is nice for high availability, but does not work with all applications or pods.