How to remove a node from a Pacemaker cluster

Removing a node from a Pacemaker cluster prevents resources from being scheduled on hardware that is being decommissioned, rebuilt, or repurposed, and keeps cluster membership clean for accurate status and fencing decisions.

In a typical Pacemaker stack, Corosync provides cluster messaging and membership, while Pacemaker stores the CIB and decides where resources run. The pcs tool coordinates these layers, updating cluster membership and Pacemaker state from a single command interface.

A safe removal evacuates resources by placing the node in standby, stops cluster services on the target host, removes the node from the remaining members’ configuration, and deauthorizes it from pcsd. Removing the wrong node or removing a node that the cluster still needs for capacity or quorum can interrupt services, so confirm redundancy and quorum behavior before proceeding.

1. Identify the exact node name as it appears in cluster status output.

   $ sudo pcs status
   Cluster name: clustername
   Cluster Summary:
     * Stack: corosync (Pacemaker is running)
     * Current DC: node-01 (version 2.1.6-6fdc9deea29) - partition with quorum
     * Last updated: Wed Dec 31 11:47:45 2025 on node-01
     * Last change:  Wed Dec 31 11:29:25 2025 by hacluster via crmd on node-01
     * 4 nodes configured
     * 7 resource instances configured
   
   Node List:
     * Online: [ node-01 node-02 node-03 node-04 ]
   
   ##### snipped #####

2. Place the node in standby to migrate resources away from it.

   $ sudo pcs node standby node-04

3. Confirm resources are no longer running on the node in the cluster status output.

   $ sudo pcs status
   Cluster name: clustername
   ##### snipped #####
   Node List:
     * Node node-04: standby
     * Online: [ node-01 node-02 node-03 ]
   ##### snipped #####

4. Stop cluster services on the node to be removed.

   $ sudo pcs cluster stop node-04
   node-04: Stopping Cluster (pacemaker)...
   node-04: Stopping Cluster (corosync)...

   Stopping a node in a small cluster can drop quorum and stop resources depending on cluster policy.

   Related: How to start and stop a Pacemaker cluster with PCS

5. Remove the node from the cluster configuration.

   $ sudo pcs cluster node remove node-04
   Destroying cluster on hosts: 'node-04'...
   node-04: Successfully destroyed cluster
   Sending updated corosync.conf to nodes...
   node-01: Succeeded
   node-02: Succeeded
   node-03: Succeeded
   node-01: Corosync configuration reloaded

   Use --force only when pcsd is not running on the removed host, because forced removal can leave the old cluster configuration on disk.

6. Deauthorize the node from pcs host authentication.

   $ sudo pcs host deauth node-04

7. Verify the node no longer appears in cluster status output.

   $ sudo pcs status
   Cluster name: clustername
   ##### snipped #####
   Node List:
     * Online: [ node-01 node-02 node-03 ]