kubeadm reset takes more than 50 seconds to retry deleting the last etcd member #91143
Labels
area/kubeadm
kind/bug
Categorizes issue or PR as related to a bug.
sig/cluster-lifecycle
Categorizes an issue or PR as relevant to SIG Cluster Lifecycle.
What happened:
When using
kubeadm reset
to delete the last control-plane node of a highly available Kubernetes cluster or the only node of a single control-plane cluster, it always take more than 50 seconds in theremove-etcd-member
phase with a lot ofFailed to remove etcd member
errors.It's impossible to remove the last/only member of an etcd cluster. There's no need to try/retry removing it as the etcd cluster will be destroyed in next phase, then a
reset
operation can finish in a few seconds, instead of nearly one minute.What you expected to happen:
kubeadm reset
doesn't need to try/retry removing the only member of an etcd cluster to avoid unnecessary errors and reduce the execution time from one minute to a few seconds.How to reproduce it (as minimally and precisely as possible):
kubeadm init
kubeadm reset
It will take more than 50 seconds in
remove-etcd-member
phase.Anything else we need to know?:
Environment:
kubectl version
): v1.18.2cat /etc/os-release
):uname -a
):The text was updated successfully, but these errors were encountered: