In some cases of k8s node failures, it can happen that a node has shut down but the pods on that node are stuck in a Terminating state. The operator tries to delete them, but since the k8s api doesn't do anything, they are still part of the redis cluster. This means that the operator will not spawn new node pods until the terminating pods are removed.
4n4nd
commented
1 year ago
In some cases of k8s node failures, it can happen that a node has shut down but the pods on that node are stuck in a Terminating state. The operator tries to delete them, but since the k8s api doesn't do anything, they are still part of the redis cluster. This means that the operator will not spawn new node pods until the terminating pods are removed.