Remove an unreachable master node from the cluster
If a master node is unreachable, we can demote it from master and then remove it from the cluster to resolve the failure. When the node is restored, we can register it with the cluster again as a non-master node.
To remove the failed node from the cluster, complete the following steps in the local management interface of the new primary master:
- Go to the Overview tab on the Cluster Configuration page.
- Under the Nodes section, select the node to remove.
- Click Delete.
- Select the Force check box to force the removal of the node even if the node cannot be reached.
- Click Yes to confirm the operation.
- Deploy the changes.
After we remove the failed node from the cluster, we might want to restart it and ultimately restore it as a cluster member. In this case, complete some additional steps. While the node is disconnected from the network, change it to a stand-alone cluster with only a single node, as described in the following steps.
- Restore the node and use its local management interface to access the Cluster Configuration page.
- Go to General tab.
- From the overview page, remove all other nodes.
- Change the Primary master IP address to 127.0.0.1.
- Save and deploy the change.
- Troubleshoot the original failure and resolve any problems.
We now join the restored appliance back in to the original cluster. This process joins the restored node to the cluster as a non-master node:
- In the local management interface of the restored appliance, go to the Overview tab on the Cluster Configuration page.
- Click Import.
- In the Join Cluster window, click Browse to select the cluster signature file of the new primary master. We can generate the cluster signature file using the local management interface of the new primary master and selecting the Export option in the Overview tab.
- Click Join to add the current appliance to the cluster.
- Deploy the changes.
Parent topic: Cluster support