Replacing a Controller Node in a Multi-Node Cluster

This section describes how to replace a controller node in a multi-node cluster. Removing a node and joining a new node is the standard process to replace a node.

Complete the following steps to replace a node within a multi-node cluster:

  1. Wipeout upgrade the controller: The replacement node should have an empty configuration and it should be using exactly the same version of the existing cluster. For instructions on the wipeout upgrade process refer to Performing a Wipeout Upgrade for a Controller Node.
  2. Remove the failed node: On the controller GUI, select Network > Data and Control Plane > Cluster. In the Control Planes list, select the node that you want to replace and click Delete to remove the node from the cluster.
  3. Join the new node: Deploy and set up the new controller node; in the Setup Wizard choose the option Join Existing Cluster.

    Joining an Existing Cluster

  4. Type the IP address of the existing cluster and enter the controller name in the Controller Name field exactly as seen in the root element of the Organization section in the Cluster page.
    Note: A typo in the Controller Name field will cause the join operation to fail. Ensure that you enter the controller name correctly, including all case-sensitive characters.

    Identifying the Controller Name

  5. Complete the Setup Wizard. The new node will join the cluster and sync the database.

Parent topic: Replacing a Controller Node