After rebooting nodes, the networkmanager configuration for interfaces is not being picked up. Instead of the defined static IP and configuration, interfaces come up as dhcp. 

4.13.4

Every Time

1. Cluster was deployed and then end customer asked for a second interface to be added and the MTU be increased to 9000. This change was implemented by RH solutions architects. We know changing the MTU post installation is not supported. This MTU change was verified as working and worked for weeks before issues were found. 
2. For a networking hardware change (new cables) the nodes were cordoned, drained, and rebooted. When the nodes came back up, the interfaces were not picking up their expected configurations. 
3. To temporarily resolve the issue, the manual commands below were ran:
```
HOST_IP="172.29.135.31"
HOST_MASK="24"
HOST_GW="172.29.135.1"
HOST_INT="eno12399np0"

ip addr add $HOST_IP/$HOST_MASK dev $HOST_INT
ip route add 0.0.0.0/0 via $HOST_GW dev $HOST_INT

touch /run/resolv-prepender-kni-conf-done
ln -s /etc/resolv.conf /var/run/NetworkManager/resolv.conf

systemctl restart kubelet
```

Node interfaces do not pick up NMstate defined configurations

Node interfaces start and use the NMstate defined configurations

Cluster must-gather: https://attachments.access.redhat.com/hydra/rest/cases/03589546/attachments/b652b2ac-5eb2-46cf-8eac-31431a615299?usePresignedUrl=true

Node sosreport: https://gss--c.vf.force.com/apex/Case_View?usePresignedUrl=true

NMstate configuration files: https://attachments.access.redhat.com/hydra/rest/cases/03589546/attachments/61bd9cf5-9e05-456b-aa20-331e47b23bee?usePresignedUrl=true

sample-01 files are previous to the reboot. I also tried using sample-02 files manually once the systems were back online, but network manager still ignores these deployments, fails on reboot, and the eno12399np0 nics are in DHCP mode (in both cases).