# Failover of a control plane node in single node/ HA Kubernetes Cluster created using kubeadm

**URL:** https://discuss.kubernetes.io/t/failover-of-a-control-plane-node-in-single-node-ha-kubernetes-cluster-created-using-kubeadm/11162
**Category:** General Discussions
**Created:** [May 30, 2020, 7:14pm UTC](https://discuss.kubernetes.io/t/failover-of-a-control-plane-node-in-single-node-ha-kubernetes-cluster-created-using-kubeadm/11162 "2020-05-30T19:14:38Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![sgarg](https://avatars.discourse-cdn.com/v4/letter/s/ad7895/32.png) [@sgarg](https://discuss.kubernetes.io/u/sgarg)
#### Post date: [May 30, 2020, 7:14pm UTC](https://discuss.kubernetes.io/t/failover-of-a-control-plane-node-in-single-node-ha-kubernetes-cluster-created-using-kubeadm/11162/1 "2020-05-30T19:14:38Z")

</div>

Hi ,

I have the following doubts:

Scenario 1: kubeadm cluster version upgrade (single control plane node cluster), when the control plane is down during upgrade.

If one of the node/pod fails, and then the master node fails (due to upgrade/ irrespective of upgrade), would the cluster have to be manually restarted and how will etcd get the state of the pods post restart?

Scenario 2 : In case of HA control plane cluster, if there are 3 control plane running, and one of them goes down (due to upgrade / irrespective of upgrade-hardware failure) .

Would the control plane node be automatically created ? If Yes, what version (new or old)?