[ClusterLabs] question about re-adding a node

[Brian Reichert]

I'm exploring some simple cluster management, and was exploring the
workflow of removing, then re-adding a node to a cluster.

I thought I understood the steps, but things are not working out
for me as I hoped.

Hopefully, someone here can provide some insight about what steps I've
skipped might be...

First, add a node:

  [root at node1 ~]# pcs cluster node add --start node3.example.com
  Disabling SBD service...
  node3.example.com: sbd disabled
  Sending remote node configuration files to 'node3.example.com'
  node3.example.com: successful distribution of the file 'pacemaker_remote
  authkey'
  node1.example.com: Corosync updated
  node2.example.com: Corosync updated
  Setting up corosync...
  node3.example.com: Succeeded
  node3.example.com: Starting Cluster (corosync)...
  Starting Cluster (pacemaker)...
  Synchronizing pcsd certificates on nodes node3.example.com...
  node3.example.com: Success
  Restarting pcsd on the nodes in order to reload the certificates...
  node3.example.com: Success

Remove that same node, seemingly without error:

  [root at node1 ~]# pcs --force cluster stop node3.example.com
  node3.example.com: Stopping Cluster (pacemaker)...
  node3.example.com: Stopping Cluster (corosync)...
  [root at node1 ~]# crm_node --force -R node3.example.com
  [root at node1 ~]# echo $?
  0

On a lark, try to re-add it:

  [root at node1 ~]# pcs cluster node add --start node3.example.com
  Error: Unable to add 'node3.example.com' to cluster: node is
  already in a cluster

Not what I was expecting; what should I be doing differently?

-- 
Brian Reichert				<reichert at numachi.com>
BSD admin/developer at large