Hi friends, In the etcd backup/restore etcd practice test, the last restore solution does not makes “kube-apiserver stop” before restoring, is that a good practice ?
Hi… I don’t think we need to stop api server. However stopping etcd is a good practice. Since the k8s cluster is configured using kubeadm you can just move the etcd static pod /etc/kubernetes/manifest/etcd-controlplane.yaml to another directory to stop it. [eg: perform your restore, stop etcd; mv /etc/kubernetes/manifest/etcd-controlplane.yaml /root, start etcd; mv /root/etcd-controlplane.yaml /etct/kubernetes/manifest]