Add NodeFeatureRules to label Nvidia NICs

This patch does the following: - Add "Nvidia NICs PCI" NFR to lable the node with `pci-15b3.present": "true"` if any of Nvidia NICs is exist - Add nfd.deployNodeFeatureRules flag and set it true by default to allow creating NFRs above - Remove the networking PCI class from the source configuration - Enhance READE documentation Signed-off-by: Waleed Mousa <waleedm@nvidia.com>
Mellanox · Jul 16, 2023 · dec8fc4 · dec8fc4
1 parent 802bb33
commit dec8fc4
Show file tree

Hide file tree

Showing 5 changed files with 33 additions and 11 deletions.
diff --git a/README.md b/README.md
@@ -57,16 +57,15 @@ NFD is used to label nodes with the following labels:
 - RDMA capability
 - GPU features*
 
+>__NOTE__: We use nodeFeatureRules to label PCI vendor and device and it's enabled by default using `nfd.deployNodeFeatureRules` flag
+
 __Example NFD worker configurations:__
 
 ```yaml
     config:
       sources:
         pci:
           deviceClassWhitelist:
-          - "02"
-          - "0200"
-          - "0207"
           - "0300"
           - "0302"
           deviceLabelFields:

diff --git a/deployment/network-operator/README.md b/deployment/network-operator/README.md
@@ -101,9 +101,19 @@ $ kubectl -n network-operator get pods
 By default the network operator
 deploys [Node Feature Discovery (NFD)](https://github.com/kubernetes-sigs/node-feature-discovery)
 in order to perform node labeling in the cluster to allow proper scheduling of Network Operator resources. If the nodes
-where already labeled by other means, it is possible to disable the deployment of NFD by setting
-`nfd.enabled=false` chart parameter.
+where already labeled by other means (either deployed from master or deployed within another deployment), it is possible to disable the deployment of NFD by setting
+`nfd.enabled=false` chart parameter and make sure that the installed version is `v0.13.2` or newer and has NodeFeatureApi enabled.
 
+##### Deploy NFD from master with NodeFeatureApi enabled
+```
+$ export NFD_NS=node-feature-discovery
+$ helm repo add nfd https://kubernetes-sigs.github.io/node-feature-discovery/charts
+$ helm repo update
+$ helm install nfd/node-feature-discovery --namespace $NFD_NS --create-namespace --generate-name --set enableNodeFeatureApi='true'
+```
+For additional information , refer to the official [NVD deployment with Helm](https://kubernetes-sigs.github.io/node-feature-discovery/v0.13/deployment/helm.html)
+
+##### Deploy Network Operator without Node Feature Discovery
 ```
 $ helm install --set nfd.enabled=false -n network-operator --create-namespace --wait network-operator mellanox/network-operator
 ```
@@ -335,6 +345,7 @@ parameters.
 | Name                                                 | Type   | Default                                  | description                                                                                                                                                 |
 |------------------------------------------------------|--------|------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------|
 | `nfd.enabled`                                        | bool   | `True`                                   | deploy Node Feature Discovery                                                                                                                               |
+| `nfd.deployNodeFeatureRules`                         | bool   | `True`                                   | deploy Node Feature Rules to label the nodes                                                                                                                |
 | `sriovNetworkOperator.enabled`                       | bool   | `False`                                  | deploy SR-IOV Network Operator                                                                                                                              |
 | `upgradeCRDs`                                        | bool   | `True`                                   | enable CRDs upgrade with helm pre-install and pre-upgrade hooks                                                                                             |
 | `sriovNetworkOperator.configDaemonNodeSelectorExtra` | object | `{"node-role.kubernetes.io/worker": ""}` | Additional nodeSelector for sriov-network-operator config daemon. These values will be added in addition to default values managed by the network-operator. |

diff --git a/deployment/network-operator/templates/nodefeaturerules.yaml b/deployment/network-operator/templates/nodefeaturerules.yaml
@@ -0,0 +1,16 @@
+{{- if .Values.nfd.deployNodeFeatureRules }}
+apiVersion: nfd.k8s-sigs.io/v1alpha1
+kind: NodeFeatureRule
+metadata:
+  name: nvidia-nics-rules
+spec:
+  rules:
+    - name: "Nvidia NICs PCI"
+      labels:
+        "pci-15b3.present": "true"
+      matchFeatures:
+        - feature: pci.device
+          matchExpressions:
+            vendor: {op: In, value: ["15b3"]}
+            class: {op: In, value: ["0200", "0207"]}
+{{- end }}
diff --git a/deployment/network-operator/values.yaml b/deployment/network-operator/values.yaml
@@ -18,6 +18,7 @@
 
 nfd:
   enabled: true
+  deployNodeFeatureRules: true
 
 psp:
   enabled: false
@@ -52,9 +53,6 @@ node-feature-discovery:
       sources:
         pci:
           deviceClassWhitelist:
-          - "02"
-          - "0200"
-          - "0207"
           - "0300"
           - "0302"
           deviceLabelFields:

diff --git a/hack/templates/values/values.template b/hack/templates/values/values.template
@@ -18,6 +18,7 @@
 
 nfd:
   enabled: true
+  deployNodeFeatureRules: true
 
 psp:
   enabled: false
@@ -52,9 +53,6 @@ node-feature-discovery:
       sources:
         pci:
           deviceClassWhitelist:
-          - "02"
-          - "0200"
-          - "0207"
           - "0300"
           - "0302"
           deviceLabelFields: