Description of problem:
When disableMultus is enabled in a HyperShift cluster, Multus is correctly not deployed by the CNO, but ovnkube-node still expects NAD. no CNI config is written, and the network remains in a NetworkNotReady state.

Version-Release number of selected component (if applicable):
Tested on OpenShift 4.20 nightly

How reproducible:
Always, when disableMultus is enabled in a HyperShift deployment using OVN-Kubernetes. 

Steps to Reproduce:

1. Create a guest cluster with --disable-multus set using the implementation here - https://github.com/openshift/hypershift/pull/6358

1. Wait for hc to start and create a worker. check the state of ovnk in HC.

1. Observe the logs

Actual results:

• ovnk pod crashes repeatedly.

• /etc/kubernetes/cni/net.d/ is empty — no CNI config is written.

• Error in logs: failed to list *v1.NetworkAttachmentDefinition: the server could not find the requested resource (get network-attachment-definitions.k8s.cni.cncf.io)

Expected results:

ovnk should not expect the NAD CRD when Multus is disabled (I believe).