How do I remove a database node from a cluster if the node is no longer accessible?

Database clustering issues will occur if the IP address of one of the clustered  database nodes changes, or if the Meeting Server is not accessible.  A ‘zombie’ node will appear in the database cluster list and the database cluster will need to be ‘rebuilt’.

To check the database cluster list, use the MMP command ‘database cluster status’.

To rebuild the database cluster, follow these steps:

  1. Use the MMP command 'database cluster remove' on all accessible nodes in the database cluster. (This step is best practise, but not required).
  2. Use  'database cluster initialize' on the server with the database contents that will used as the one and only database instance. This will be the primary database.
  3. Use  'database cluster join <hostname/IP address>' on the other two servers to create the database cluster. This command will destroy the contents of any previous database overwriting it with the contents of the primary database created in step 2.
  4. Use  'database cluster connect <hostname/IP address>' to connect a Call Bridge to the database cluster.

Note: From version 2.7, these commands will not run without valid certificates, keys and CA certificates uploaded to the database clients and servers.

Last update: