Change nncp to match new label added to nodes

[tssala23]

There are two types of worker nodes, non-GPUs and GPUs. The GPUs on this cluster only have one interface, this means that networking for these nodes are slightly different. This change makes it so this nncp is only applied to the non-GPU nodes.
Labels were added to the non-gpu nodes:
kubectl label node <non-gpu-wrk> gpu=false --as system:admin

[larsks]

Are we running the nvidia node discovery operator on this cluster? I think we should be if we have GPU nodes.

[jtriley]

Are we running the nvidia node discovery operator on this cluster? I think we should be if we have GPU nodes.

@tssala23 To get the nvidia-gpu and node-feature-discovery operators working on a cluster:

1. Include node-feature-discovery bundle in your cluster-scope overlay's top-level kustomization.yaml
2. Include nvidia-gpu-operator bundle in your cluster-scope overlay's top-level kustomization.yaml
3. Create a cluster overlay in the nfd-discovery top-level directory
4. Create a cluster overlay in the nvidia-gpu-operator top-level directory
5. Add a node-feature-discovery argocd app to to the nerc-ocp-apps repo
6. Add an nvidia-gpu-operator argocd app to the nerc-ocp-apps repo

For the argocd apps, see:

https://github.com/OCP-on-NERC/nerc-ocp-apps/blob/main/clusters/nerc-ocp-test/kustomization.yaml#L40-L54

[joachimweyl]

@tssala23 what are the next steps on this?

[tssala23]

@joachimweyl This can be put in the icebox for now. We implemented the change without using nncp and the nncp dont currently work properly though I do want to come back to it at some point, Not important now.