You can use repmgr on a running node on your cluster to unregister a failed or unreachable standby node.

Note: For the primary node to function normally, at least one standby node must always be running.


To unregister a standby node that is not running, you must provide the node ID. See Check the Connectivity Status of a Database High Availability Cluster.


  1. Log in or SSH as root to the OS of any of the running nodes of the cluster.
  2. Change the user to postgres.
    sudo -i -u postgres
  3. Unregister the failed or unreachable node.
    /opt/vmware/vpostgres/current/bin/repmgr standby unregister -f /opt/vmware/vpostgres/current/etc/repmgr.conf --node-id=ID


Unregistering the node removes the node information from the repmgr metadata.