Windows Machine Config Operator does not set the NodeIP when the node IP is provided through a "windows-instances" configmap. 

This is an issue on Nutanix (we have a client waiting for Windows support on Nutanix) and here is a gist with the same issue, about a year ago : https://github.com/therevoman/nate-revo-blog/issues/4

The fix I found is the same as detailled in the gist : add the parameter "--node-ip=<HOST IP>" to the kubelet service.

Could we make this the default behavior when boostrapping the node based on an IP in BYOH mode with the configmap ? I think this is a quick and easy fix, which might help in other situations.

OpenShift 4.12.8 and WMCO 7.0.1

Every time I tried on Nutanix I had the issue.

1. Create a template the same way you would on VSphere (windows 2022 EN-US)
2. Create a new VM from the template, add a PTR record, change the machine hostname and add the node IP to the "windows-instances" configmap as specified in the BYOH doc of the WMCO.
3. Watch the operator logs and see it timeout after the kube-proxy service step (at the end of bootstraping)
4. See that the node as been added to the list of nodes but is not schedulable as it does not have the right IP (eventhough the windows host only has one interface) 

Node is added with the wrong IP (VXLan IP instead of real interface IP) in the kubelet service.

Have the node in the cluster with the good IP right away.

I have sometimes also encountered issues with the "C:\Temp\netwotk-configuration.ps1" script, which did not work when called by wicd at first but worked afterwards if manually called beforehand. I think it might be related to the module imported but not really useable by the script.