Worker nodes tainted and NotReady after cluster upgrade

Summary

Cluster setup:

• Six node cluster where the first machine was configured with a launch configuration.
• The other five nodes join the first node through a launch configuration also.
• The last three nodes join as worker nodes.

Upgrading the cluster from 1.29.4 to 1.30.1 resulted in the worker nodes stuck in NotReady state.

Mitigation steps attempted (failed):

• Remove taints (NoSchedule, NoExecute)
• Re-apply the launch configuration on the worker nodes

What did work:

• 'leave' on the worker nodes
• 'delete node' on the control plane
• 'join' with the token generated from 'add-node'

What Should Happen Instead?

The worker nodes should have successfully reconnected to the cluster just like the control plane nodes.

Reproduction Steps

1. Set up node 1 with a launch configuration.
2. Join two more control plane nodes with a launch configuration.
3. Join three worker nodes with a launch configuration.
4. Upgrade the cluster nodes in order (CP1, CP2, CP3, WN1, WN2, WN3)

Inspection from WN1 inspection.tar.gz