When applying NodeNetworkConfigurationPolicy to create a Shim interface, it initially is configured correctly & everything works as expected. But as soon as the Node is rebooted, the Node is stuck in NotReady state, since configure-ovs.sh fails to recreate OVNKubernetes. 

The NodeNetworkConfigurationPolicy is based on the ODF multus ceph-public-net example here: https://docs.redhat.com/en/documentation/red_hat_openshift_data_foundation/4.17/html/planning_your_deployment/network-requirements_rhodf#multus-examples_rhodf

This makes it currently impossible to have a stable multus deployment of ODF.

CoreOS: 418.94.202502100215-0 / nmstate: kubernetes-nmstate-operator.4.18.0-202412040208

I can reproduce it every time with OCP 4.18, OCP 4.17 & OCP 4.16. It works correctly with OCP 4.15. It is simple to reproduce: you just need a new NIC, apply the attached YAML on it & reboot the node.   

1. Select a Node, e.g. worker-001
2. Add a new NIC, e. g. ence400
3. Apply the NMState YAML attached to this issue
4. Reboot

At first, everything works. But after Node reboot, the Node remains in NotReady state: configure-ovs.sh script does not recreate OVNKubernetes.

Node should be Ready after reboot.

I can observe this issue since OCP 4.16    

Please note, that with simpler NodeNetworkConfigurationPolicy I have no issues, e.g. by simply configuring the state or ipv4 configuration of a NIC.