Merge pull request #512 from e0ne/rel-23.4.0-rc.1

Release Network Operator 23.4.0-rc.1

Author: adrianchiris

deployment/network-operator/Chart.yaml

@@ -1,8 +1,8 @@
 apiVersion: v2
 name: network-operator
-version: 23.4.0-beta.4
+version: 23.4.0-rc.1
 kubeVersion: '>= 1.21.0'
-appVersion: v23.4.0-beta.4
+appVersion: v23.4.0-rc.1
 description: Nvidia network operator
 type: application
 keywords:

hack/templates/samples/mellanox.com_v1alpha1_nicclusterpolicy.template

@@ -20,12 +20,23 @@ spec:
     image: {{ .Mofed.Image }}
     repository: {{ .Mofed.Repository }}
     version: {{ .Mofed.Version }}
+    livenessProbe:
+      initialDelaySeconds: 30
+      periodSeconds: 30
+    readinessProbe:
+      initialDelaySeconds: 10
+      periodSeconds: 30
+    startupProbe:
+      initialDelaySeconds: 10
+      periodSeconds: 20
+    terminationGracePeriodSeconds: 300
     upgradePolicy:
       autoUpgrade: false
       drain:
         deleteEmptyDir: true
         enable: true
         force: true
+        podSelector: ""
         timeoutSeconds: 300
       maxParallelUpgrades: 1
   rdmaSharedDevicePlugin:

hack/templates/values/values.template

@@ -24,6 +24,9 @@ psp:
 
 sriovNetworkOperator:
   enabled: false
+  # inject additional values to nodeSelector for config daemon
+  configDaemonNodeSelectorExtra:
+    node-role.kubernetes.io/worker: ""
 
 # Node Feature discovery chart related values
 node-feature-discovery:
@@ -169,9 +172,9 @@ ofedDriver:
     # global switch for automatic upgrade feature
     # if set to false all other options are ignored
     autoUpgrade: false
-    # how many nodes can be upgraded in parallel
+    # how many nodes can be upgraded in parallel (default: 1)
     # 0 means no limit, all nodes will be upgraded in parallel
-    maxParallelUpgrades: 0
+    maxParallelUpgrades: 1
     # options for node drain (`kubectl drain`) before the driver reload
     # if auto upgrade is enabled but drain.enable is false,
     # then driver POD will be reloaded immediately without