feat: Create gpu-provisioner helm values template

Makefile

@@ -152,21 +152,6 @@ vet: ## Run go vet against code.
 lint: $(GOLANGCI_LINT)
 	$(GOLANGCI_LINT) run -v
 
-## --------------------------------------
-## Release
-## To create a release, run `make release VERSION=x.y.z`
-## --------------------------------------
-.PHONY: release-manifest
-release-manifest:
-	@sed -i -e 's/^VERSION ?= .*/VERSION ?= ${VERSION}/' ./Makefile
-	@sed -i -e "s/version: .*/version: ${IMG_TAG}/" ./charts/gpu-provisioner/Chart.yaml
-	@sed -i -e "s/appVersion: .*/appVersion: ${IMG_TAG}/" ./charts/gpu-provisioner/Chart.yaml
-	@sed -i -e "s/tag: .*/tag: ${IMG_TAG}/" ./charts/gpu-provisioner/values.yaml
-	@sed -i -e 's/gpu-provisioner: .*/gpu-provisioner:${IMG_TAG}/' ./charts/gpu-provisioner/README.md
-	git checkout -b release-${VERSION}
-	git add ./Makefile ./charts/gpu-provisioner/Chart.yaml ./charts/gpu-provisioner/values.yaml ./charts/gpu-provisioner/README.md
-	git commit -s -m "release: update manifest and helm charts for ${VERSION}"
-
 ## --------------------------------------
 ## Tests
 ## --------------------------------------
@@ -188,3 +173,19 @@ e2etests: ## Run the e2e suite against your local cluster
 		--ginkgo.timeout=${TEST_TIMEOUT} \
 		--ginkgo.grace-period=3m \
 		--ginkgo.vv
+
+## --------------------------------------
+## Release
+## To create a release, run `make release VERSION=x.y.z`
+## --------------------------------------
+.PHONY: release-manifest
+release-manifest:
+	@sed -i -e 's/^VERSION ?= .*/VERSION ?= ${VERSION}/' ./Makefile
+	@sed -i -e "s/version: .*/version: ${IMG_TAG}/" ./charts/gpu-provisioner/Chart.yaml
+	@sed -i -e "s/appVersion: .*/appVersion: ${IMG_TAG}/" ./charts/gpu-provisioner/Chart.yaml
+	@sed -i -e "s/tag: .*/tag: ${IMG_TAG}/" ./charts/gpu-provisioner/values.yaml
+	@sed -i -e 's/gpu-provisioner: .*/gpu-provisioner:${IMG_TAG}/' ./charts/gpu-provisioner/README.md
+	@sed -i -e 's/CHART_VERSION=.*/CHART_VERSION=${IMG_TAG}/' ./charts/gpu-provisioner/README.md
+	git checkout -b release-${VERSION}
+	git add ./Makefile ./charts/gpu-provisioner/Chart.yaml ./charts/gpu-provisioner/values.yaml ./charts/gpu-provisioner/README.md
+	git commit -s -m "release: update manifest and helm charts for ${VERSION}"

README.md

@@ -8,15 +8,15 @@ gpu-Provisioner is an [Azure Karpenter provider](https://github.com/Azure/karpen
 It implements the cloud provider interfaces to realize the following abstraction:
 `machine` -> `AKS agent pool` (with vmss and a hard limit of VM count to 1)
 
-```
-VERSION=v0.2.0 make docker-build
-make az-identity-perm
-make az-patch-helm
-helm install gpu-provisioner /charts/gpu-provisioner --namespace gpu-provisioner --create-namespace
-make az-federated-credential
-```
-You should have a running controller in `gpu-provisioner` namespace.
+## Prerequisites
+- An Azure subscription.
+- An AKS cluster with [OIDC](https://learn.microsoft.com/en-us/azure/aks/use-oidc-issuer) addon installed. Please refer to the [Karpenter installation guide](https://karpenter.sh/docs/installation/) for more details.
+- 
+## Install gpu-provisioner
 
+Please check the installation guidance [here](./charts/gpu-provisioner/README.md).
+
+```shell
 ## How to test
 After deploying the controller successfully, one can apply the yaml in `/examples` to create a machine CR. A real node will be created and added to the cluster by the controller.
 

charts/gpu-provisioner/README.md

@@ -9,7 +9,20 @@ A Helm chart for gpu-provisioner
 To install the chart with the release name `gpu-provisioner`:
 
 ```bash
-helm install gpu-provisioner ./charts/gpu-provisioner --namespace gpu-provisioner --create-namespace
+export CHART_VERSION=0.2.0
+export CLUSTER_NAME=my-cluster
+export AZURE_RESOURCE_GROUP=my-rg
+export AZURE_SUBSCRIPTION_ID=my-subscription-id
+export MSI_NAME=gpuIdentity
+
+az identity create --name $MSI_NAME --resource-group $CLUSTER_NAME
+
+./hack/deploy/configure-helm-values.sh $CLUSTER_NAME $AZURE_RESOURCE_GROUP $MSI_NAME
+
+helm install gpu-provisioner \
+      https://github.com/Azure/gpu-provisioner/raw/gh-pages/charts/gpu-provisioner-$CHART_VERSION.tgz \
+      --values gpu-provisioner-values.yaml --namespace gpu-provisioner --create-namespace --wait
+make az-federated-credential
 ```
 
 ## Values
@@ -47,7 +60,7 @@ helm install gpu-provisioner ./charts/gpu-provisioner --namespace gpu-provisione
 | podLabels                          | object | `{}`                                                                                                                                                                                   | Additional labels for the pod.                                                                                         |
 | podSecurityContext                 | object | `{"fsGroup":1000}`                                                                                                                                                                     | SecurityContext for the pod.                                                                                           |
 | priorityClassName                  | string | `"system-cluster-critical"`                                                                                                                                                            | PriorityClass name for the pod.                                                                                        |
-| replicas                           | int    | `2`                                                                                                                                                                                    | Number of replicas.                                                                                                    |
+| replicas                           | int    | `1`                                                                                                                                                                                    | Number of replicas.                                                                                                    |
 | revisionHistoryLimit               | int    | `10`                                                                                                                                                                                   | The number of old ReplicaSets to retain to allow rollback.                                                             |
 | serviceAccount.annotations         | object | `{}`                                                                                                                                                                                   | Additional annotations for the ServiceAccount.                                                                         |
 | serviceAccount.create              | bool   | `true`                                                                                                                                                                                 | Specifies if a ServiceAccount should be created.                                                                       |

gpu-provisioner-values-template.yaml

@@ -0,0 +1,23 @@
+
+replicas: 1 # for better debugging experience
+controller:
+  env:
+    # Azure client settings
+    - name: ARM_SUBSCRIPTION_ID
+      value: ${AZURE_SUBSCRIPTION_ID}
+    - name: LOCATION
+      value: ${AZURE_LOCATION}
+    - name: AZURE_CLUSTER_NAME
+      value: ${CLUSTER_NAME}
+    - name: AZURE_NODE_RESOURCE_GROUP
+      value: ${AZURE_RESOURCE_GROUP_MC}
+    - name: ARM_RESOURCE_GROUP
+      value: ${AZURE_RESOURCE_GROUP}
+    - name: LEADER_ELECT # disable leader election for better debugging experience
+      value: "false"
+    - name: E2E_TEST_MODE
+      value: "false"
+
+workloadIdentity:
+  clientId: ${GPU_PROVISIONER_USER_ASSIGNED_CLIENT_ID}
+  tenantId: ${AZURE_TENANT_ID}

hack/deploy/configure-helm-values.sh

@@ -0,0 +1,34 @@
+// https://github.com/Azure/karpenter-provider-azure/blob/2beb773cbd3134eeabb8c96b72a130b86b1a91e1/hack/deploy/configure-values.sh
+
+#!/usr/bin/env bash
+set -euo pipefail
+
+# This script interrogates the AKS cluster and Azure resources to generate
+# the gpu-provisioner-values.yaml file using the gpu-provisioner-values-template.yaml file as a template.
+
+if [ "$#" -ne 3 ]; then
+    echo "Usage: $0 <cluster-name> <resource-group> <gpu-provisioner-user-assigned-identity-name>"
+    exit 1
+fi
+
+echo "Configuring gpu-provisioner-values.yaml for cluster $1 in resource group $2 ..."
+
+CLUSTER_NAME=$1
+AZURE_RESOURCE_GROUP=$2
+AZURE_GPU_PROVISIONER_USER_ASSIGNED_IDENTITY_NAME=$3
+
+AKS_JSON=$(az aks show --name "$CLUSTER_NAME" --resource-group "$AZURE_RESOURCE_GROUP")
+AZURE_LOCATION=$(jq -r ".location" <<< "$AKS_JSON")
+AZURE_RESOURCE_GROUP_MC=$(jq -r ".nodeResourceGroup" <<< "$AKS_JSON")
+AZURE_TENANT_ID=$(az account show |jq -r ".tenantId")
+
+
+GPU_PROVISIONER_USER_ASSIGNED_CLIENT_ID=$(az identity show --resource-group "${AZURE_RESOURCE_GROUP}" --name "${AZURE_GPU_PROVISIONER_USER_ASSIGNED_IDENTITY_NAME}" --query 'clientId' -otsv)
+
+export CLUSTER_NAME AZURE_LOCATION AZURE_RESOURCE_GROUP_MC GPU_PROVISIONER_USER_ASSIGNED_CLIENT_ID AZURE_TENANT_ID
+
+# get gpu-provisioner-values-template.yaml, if not already present (e.g. outside of repo context)
+if [ ! -f gpu-provisioner-values-template.yaml ]; then
+    curl -sO https://raw.githubusercontent.com/Azure/gpu-provisioner/main/gpu-provisioner-values-template.yaml
+fi
+yq '(.. | select(tag == "!!str")) |= envsubst(nu)' gpu-provisioner-values-template.yaml > gpu-provisioner-values.yaml