This guide describes a recommended workflow for a previously decommissioned node to rejoin the same cluster.

  1. Cleanup Portworx metadata on the node
    • Portworx 1.3 and higher: /opt/pwx/bin/pxctl sv node-wipe --all
    • Portworx 1.2: rm -rf /etc/pwx
  2. Restart Portworx pod on the node by cleaning up labels.
    • kubectl label nodes <node> px/service- --overwrite
    • kubectl label nodes <node> px/enabled- --overwrite

In above command, <node> is the node which is rejoining the cluster. After above steps, Portworx will start on this node and will join the cluster back as a new node.