1. Proposed title of this feature request
Leave Nodes cordoned after upgrade

2. What is the nature and description of the request?
The MachineConfigOperator is used to manage on-disk state for OpenShift Nodes. We have a client that has marked some of their Nodes as Unschedulable (cordonned).

They have performed changes to their MachineConfig and this has created a new rendered-machine-config which is applied to the MachineConfigPool and rolled out to all of the Nodes in the MCP. (This is possible because the maxUnavailable > currently cordonned)

When the upgrade completes, the previously-cordonned Nodes are now schedulable.

This is undesired by the customer and would prefer the MCO know whether it cordonned the Node for the upgrade or not.

3. Why does the customer need this? (List the business requirements here)
The customer has marked these Nodes as unschedulable as they do not want workloads deployed on these Nodes yet but the MCO is overwritting this configuration.

4. List any affected packages or components.
MachineConfigOperator