[DOC] Update Full DeviceConfig Example for release v1.1.0 (#54)

* Update Full DeviceConfig Example * Update docs/fulldeviceconfig.rst Co-authored-by: Matt Elliott <[email protected]> * Update docs/fulldeviceconfig.rst Co-authored-by: Matt Elliott <[email protected]> * Update docs/fulldeviceconfig.rst Co-authored-by: Matt Elliott <[email protected]> * Update docs/fulldeviceconfig.rst Co-authored-by: Matt Elliott <[email protected]> * Apply more suggestions --------- Co-authored-by: Matt Elliott <[email protected]>
ROCm · Feb 4, 2025 · b870c8d · b870c8d
1 parent c2bde6c
commit b870c8d
Showing 1 changed file with 70 additions and 16 deletions.
diff --git a/docs/fulldeviceconfig.rst b/docs/fulldeviceconfig.rst
@@ -28,19 +28,26 @@ Below is an example of a full DeviceConfig CR that can be used to install the AM
 .. code-block:: yaml
   
     apiVersion: amd.com/v1alpha1 
-    kind: DeviceConfig #New Custom Resource Definition used by the GPU Operator
+    kind: DeviceConfig # New Custom Resource Definition used by the GPU Operator
     metadata:
-      # Name of the DeviceConfig CR. Note that the name of device plugin, node-labeller and metric-explorter pods will be prefixed with 
-      name: gpu-operator 
-      namespace: kube-amd-gpu # Namespace for the GPU Operator and it's components
+      # Name that will prefix device plugin, node-labeller and metrics-exporter pods
+      name: gpu-operator
+      # Namespace where the GPU Operator and its components will run
+      namespace: kube-amd-gpu
     spec: 
       ## AMD GPU Driver Configuration ##
       driver:
-        # Set to false to skip driver installation to use inbox or pre-installed driver on worker nodes
-        # Set to True to enable operator to install out-of-tree amdgpu kernel module
+        # Set to false to use existing in-tree/pre-installed driver
+        # Set to true to install out-of-tree amdgpu kernel module
+        # Default: true
         enable: false 
-        blacklist: false # Set to true to blacklist the amdgpu kernel module which is required for installing out-of-tree driver
-        # Specify your repository to host driver image
+        # Set blacklist to true to blacklist the inbox / pre-installed amdgpu kernel module
+        # Required when spec.driver.enable is true
+        # GPU Worker node reboot is required to apply the blacklist
+        blacklist: false
+        # Specify the out-of-tree amdgpu driver version you want to install that coincides with a ROCm version number
+        version: "6.3"
+        # Specify your repository URL to host driver image for out-of-tree amdgpu kernel module
         # DO NOT include the image tag as AMD GPU Operator will automatically manage the image tag for you
         image: docker.io/username/repo
         # (Optional) Specify the credential for your private registry if it requires credential to get pull/push access
@@ -49,23 +56,70 @@ Below is an example of a full DeviceConfig CR that can be used to install the AM
         # Make sure you created the secret within the namespace that KMM operator is running
         imageRegistrySecret:
           name: mysecret
+        # (Optional) Specify your image registry's TLS config
         imageRegistryTLS: 
           insecure: False # If True, check for the container image using plain HTTP
           insecureSkipTLSVerify: False # If True, skip any TLS server certificate validation (useful for self-signed certificates)
-        version: "6.3" # Specify the driver version you would like to be installed that coincides with a ROCm version number
+        # (Optional) Specify configuration to sign the driver image
+        # Will be used when there is no pre-compiled driver image 
+        # and operator is building + signing driver image in one shot within cluster
+        # necessary for secure boot enabled system
+        imageSign:
+          # the private key used to sign kernel modules within image
+          keySecret:
+            name: my-key-secret
+          # the public key used to sign kernel modules within image
+          certSecret:
+            name: my-cert-secret
       ## AMD K8s Device Plugin Configuration ##
       devicePlugin: 
         # (Optional) Specifying image names are optional. Default image names for shown here if not specified.
         devicePluginImage: rocm/k8s-device-plugin:latest # Change this to trigger metrics exporter upgrade on CR update
         nodeLabellerImage: rocm/k8s-device-plugin:labeller-latest # Change this to trigger metrics exporter upgrade on CR update
+        # (Optional) Specify image registry secret to pull device plugin and node labeller images if needed. 
+        imageRegistrySecret:
+          name: my-deviceplugin-image-secret
+        # (Optional) Enable or disable node labeller, default value is true
+        enableNodeLabeller: true
       ## AMD GPU Metrics Exporter Configuration ##
-      metricsExporter: 
-        enable: False # False by Default. Set to True to enable the Metrics Exporter 
-        serviceType: ClusterIP # ServiceType used to expose the Metrics Exporter endpoint. Can be either `ClusterIp` or `NodePort`.
-        port: 5000 # Used to specify Port the Metrics Exporter service is exposed on when using ClusterIP serviceType
-        nodePort: 32500 # Used instead of `port` when using NodePort as the serviceType. The port number must be between 30000-32767
-        # (Optional) Specifying metrics exporter image is optional. Default imagename shown here if not specified.
-        image: rocm/device-metrics-exporter:latest # Change this to trigger metrics exporter upgrade on CR update
+      metricsExporter:
+        # Enable metrics collection and exposure (Default: false)
+        enable: False
+        # Service type for metrics endpoint exposure
+        # Values: ClusterIP, NodePort
+        # Default: ClusterIP
+        serviceType: ClusterIP
+        # Port for metrics endpoint when using ClusterIP
+        # Default: 5000
+        port: 5000
+        # Port for metrics endpoint when using NodePort
+        # Valid range: 30000-32767
+        # Default: 32500
+        nodePort: 32500
+        # Container image for metrics exporter
+        # Default: rocm/device-metrics-exporter:latest
+        image: rocm/device-metrics-exporter:latest
+        # Private registry credentials (optional)
+        imageRegistrySecret:
+          name: exporter-image-pull-secret
+        # Custom metrics exporter configuration (optional)
+        config:
+          name: exporter-configmap
+        # RBAC Proxy Configuration for secure metrics endpoint access (optional)
+        rbacConfig:
+          # Enable RBAC authentication proxy (Default: false)
+          # When enabled, provides authentication and authorization for metrics endpoint        
+          enable: false
+          # RBAC proxy container image
+          # Default: quay.io/brancz/kube-rbac-proxy:v0.18.1          
+          image: "quay.io/brancz/kube-rbac-proxy:v0.18.1"
+          # TLS configuration for metrics endpoint
+          # Set true to disable HTTPS          
+          disableHttps: false
+          # TLS certificate configuration
+          # Default: Auto-generated self-signed certificates         
+          secret:
+            name: my-kube-rbac-proxy-cert
         # If specifying a node selector here, the metrics exporter will only be deployed on nodes that match the selector
         # See Item #6 on https://dcgpu.docs.amd.com/projects/gpu-operator/en/latest/knownlimitations.html for example usage
         selector: