Skip to main content

Remove Node from Cluster

Introduction

These are the steps if a node needs to be removed entirely from the cluster and function as a single node

Steps

  1. Remove or mitigate all VMs and containers from nodes that will be decommissioned
  2. Create a ZFS snapshot of rpool and rpool/data on the single node
  3. Delete the nodes that will be decommissioned
  4. Remove the cluster config
  5. (Optional) Disable cluster/HA services

Snapshot

zfs snapshot rpool@<date>
zfs snapshot rpool/data@<date>

Remove the decommissioned nodes from the single node

pvecm delnode <nodename>

Run the following to  delete the cluster and create a single node

It may be best to run this line by line to see the output.

systemctl stop pve-cluster corosync
pmxcfs -l
rm /etc/corosync/*
rm /etc/pve/corosync.conf
killall pmxcfs
systemctl start pve-cluster 

(Optional) Disable cluster/HA services

systemctl disable --now pve-ha-crm pve-ha-lrm corosync.service