A master node is Offline due to one of the following reasons maintenance, OS crash, hardware failure. This article helps to replace a master node in a bare metal multi-master cluster.
- Platform9 Managed Kubernetes - All Version
- Enable Advanced Remote Support with sudo permissions as the new master node can go offline while being added to the cluster and ARS helps in troubleshooting such issues. Here's the article to assist you with it.
- Ensure that we have an etcd quorum by checking the etcd cluster's health.
# etcdctl --cert-file /etc/pf9/kube.d/certs/etcdctl/etcd/request.crt --key-file /etc/pf9/kube.d/certs/etcdctl/etcd/request.key --ca-file /etc/pf9/kube.d/certs/etcdctl/etcd/ca.crt cluster-health
- Once the above command returns a Healthy status, go to the Clusters view in the Clarity UI->Select the cluster from which you want to detach the master -> Click on Detach option
- Select the master node you would like to Detach and click on Detach Nodes.
- Check the following output to ensure that the older master node has been detached
# kubectl get nodes -w
- Once it's detached and the cluster can be seen in a Connected state in the UI, make sure that you run the pf9-hostagent installer on the new master node that you want to add to the existing cluster and you have authorized it as a Node.
- Now, click on attach Node to add a new master in the existing cluster.
- After a while, the new master should be added to the cluster.