How to Safely Shut Down and Restart a Kubernetes Cluster

Introduction

During routine Kubernetes cluster maintenance you may need to temporarily shut down or restart the cluster. This article explains how to safely shut down a K8s cluster and restart it.

Typical Node Maintenance

Shutting down a K8s cluster is risky; you must understand the consequences. First back up applications, custom resources (CRDs), and etcd, then consider draining nodes instead of rebooting the whole cluster. Example commands: $ kubectl get nodes Then drain a node: $ kubectl drain <node name> After a successful drain you can power off the node or, if you keep it, uncordon it:

kubectl uncordon <node name>

Preparation Before Cluster Shutdown

Backup is the most important step. Ensure SSH password‑less login, back up applications, custom resources, and etcd.

SSH password‑less login between hosts

Application backup

Custom resource backup

Etcd backup

Shutting Down the Kubernetes Cluster

Reminder Back up data and applications before shutting down; the method described can shut down the cluster smoothly, but data corruption is still possible.

Get the list of nodes: k8snodes=$(kubectl get nodes -o name) Then shut down each node, for example:

for node in ${k8snodes[@]}
do
    echo "==== Shut down $node ===="
    ssh $node sudo shutdown -h 1
done

Note: SSH password‑less login must be configured.

Kubernetes Cluster Restart

After restart, verify node and core component status.

$ kubectl get nodes -o wide
$ kubectl get svc -n kube-system
$ kubectl get pod -n kube-system

Cluster Restart Pitfalls

Common issues include etcd data corruption, network errors, or application failures after nodes are up. Always have multiple backups to meet RTO.

References

Kasten practical series backup K8s cloud‑native applications

Kasten K10 series 02 – backup Kubernetes etcd database