Enable NIC Feature Discovery in NVIDIA Network Operator

💡 Quick Answer: Enable NIC Feature Discovery by adding spec.nicFeatureDiscovery.enabled: true in NicClusterPolicy, then verify the daemon runs and node labels are present with oc get node --show-labels.

NIC Feature Discovery adds hardware capability labels to nodes so scheduling and network policies can target compatible hosts.

Enable in NicClusterPolicy

apiVersion: mellanox.com/v1alpha1
kind: NicClusterPolicy
metadata:
  name: nic-cluster-policy
spec:
  nicFeatureDiscovery:
    enabled: true
    repository: nvcr.io/nvidia/mellanox
    image: nic-feature-discovery
    version: network-operator-v25.10.0

Apply it:

oc apply -f nic-cluster-policy.yaml

Verify DaemonSet and Pods

oc -n nvidia-network-operator get ds | grep -i feature
oc -n nvidia-network-operator get pods -o wide | grep -i nic-feature

Verify Labels on Nodes

oc get node -o json | jq '.items[].metadata.labels'

Look for labels like:

• feature.node.kubernetes.io/pci-15b3.present=true
• feature.node.kubernetes.io/pci-15b3.sriov.capable=true
• feature.node.kubernetes.io/rdma.capable=true

Troubleshooting

• No daemonset created: confirm nicFeatureDiscovery.enabled: true is set under spec.
• No new labels: verify pods are running on target worker nodes.
• Labels exist but scheduling fails: validate node selectors in your SR-IOV and workload manifests.