Merge pull request #98 from angudadevops/ofed_fix

Anurag Guda · web-flow · commit 97d373818764 · 2025-02-26T13:55:52.000-06:00
Network Operator NIC cluster Policy Ofed fix
diff --git a/playbooks/cns_values.yaml b/playbooks/cns_values.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_12.0.yaml b/playbooks/cns_values_12.0.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_12.1.yaml b/playbooks/cns_values_12.1.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_12.2.yaml b/playbooks/cns_values_12.2.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_12.3.yaml b/playbooks/cns_values_12.3.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_13.0.yaml b/playbooks/cns_values_13.0.yaml
@@ -61,10 +61,12 @@ ngc_registry_email: ""
 ngc_registry_username: "$oauthtoken"
 
 # Network Operator Values
-## If the Network Operator is yes then make sure enable_rdma as well yes
+## If the Network Operator is yes then make sure enable_rdma as well yesgit br
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_13.1.yaml b/playbooks/cns_values_13.1.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_13.2.yaml b/playbooks/cns_values_13.2.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/cns_values_14.0.yaml b/playbooks/cns_values_14.0.yaml
@@ -65,6 +65,8 @@ ngc_registry_username: "$oauthtoken"
 enable_network_operator: no
 ## Enable RDMA yes for NVIDIA Certification
 enable_rdma: no
+## Enable for MLNX-OFED Driver Deployment
+deploy_ofed: no
 
 # Prxoy Configuration
 proxy: no
diff --git a/playbooks/files/network-operator-value.yaml b/playbooks/files/network-operator-value.yaml
@@ -0,0 +1,3 @@
+nfd:
+  enabled: true
+  deployNodeFeatureRules: true
diff --git a/playbooks/files/nic-cluster-policy.yaml b/playbooks/files/nic-cluster-policy.yaml
@@ -0,0 +1,72 @@
+apiVersion: mellanox.com/v1alpha1
+kind: NicClusterPolicy
+metadata:
+  name: nic-cluster-policy
+spec:
+{% if deploy_ofed %}
+  ofedDriver:
+    readinessProbe:
+      initialDelaySeconds: 10
+      periodSeconds: 30
+    forcePrecompiled: false
+    terminationGracePeriodSeconds: 300
+    livenessProbe:
+      initialDelaySeconds: 30
+      periodSeconds: 30
+    env:
+      - name: UNLOAD_STORAGE_MODULES
+        value: 'true'
+    upgradePolicy:
+      autoUpgrade: true
+      drain:
+        deleteEmptyDir: true
+        enable: true
+        force: true
+        timeoutSeconds: 300
+        podSelector: ''
+      maxParallelUpgrades: 1
+      safeLoad: false
+      waitForCompletion:
+        timeoutSeconds: 0
+    startupProbe:
+      initialDelaySeconds: 10
+      periodSeconds: 20
+    image: doca-driver
+    repository: nvcr.io/nvidia/mellanox
+    version: 24.10-0.7.0.0-0
+{% endif %}
+  rdmaSharedDevicePlugin:
+    image: k8s-rdma-shared-dev-plugin
+    repository: ghcr.io/mellanox
+    version: v1.5.2
+    # The config below directly propagates to k8s-rdma-shared-device-plugin configuration.
+    # Replace 'devices' with your (RDMA capable) netdevice name.
+    config: |
+      {
+        "configList": [
+          {
+            "resourceName": "rdma_shared_device_a",
+            "rdmaHcaMax": 63,
+            "selectors": {
+              "vendors": ["15b3"],
+              "ifNames": ["ens2f0"]
+            }
+          }
+        ]
+      }
+  secondaryNetwork:
+    cniPlugins:
+      image: plugins
+      repository: ghcr.io/k8snetworkplumbingwg
+      version: v1.5.0
+      imagePullSecrets: []
+    multus:
+      image: multus-cni
+      repository: ghcr.io/k8snetworkplumbingwg
+      version: v4.1.0
+      imagePullSecrets: []
+    ipamPlugin:
+      image: whereabouts
+      repository: ghcr.io/k8snetworkplumbingwg
+      version: v0.7.0
+      imagePullSecrets: []
diff --git a/playbooks/operators-install.yaml b/playbooks/operators-install.yaml

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,3 @@`
	`1`	`+nfd:`
	`2`	`+ enabled: true`
	`3`	`+ deployNodeFeatureRules: true`