[history server][e2e] Add Azure Blob Storage E2E test for History Server

historyserver/config/historyserver-azureblob.yaml

@@ -0,0 +1,51 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: historyserver
+  labels:
+    app: historyserver
+spec:
+  selector:
+    app: historyserver
+  ports:
+  - protocol: TCP
+    name: http
+    port: 30080
+    targetPort: 8080
+  type: ClusterIP
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: historyserver-demo
+  labels:
+    app: historyserver
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: historyserver
+  template:
+    metadata:
+      labels:
+        app: historyserver
+    spec:
+      serviceAccountName: historyserver
+      containers:
+      - name: historyserver
+        env:
+        - name: AZURE_STORAGE_CONNECTION_STRING
+          value: "DefaultEndpointsProtocol=http;AccountName=devstoreaccount1;AccountKey=Eby8vdM02xNOcqFlqUwJPLlmEtlCDXJ1OUzFT50uSRZ6IFsuFq2UVErCz4I6tq/K1SZFPTOtr/KBHBeksoGMGw==;BlobEndpoint=http://azurite-service.azurite-dev.svc.cluster.local:10000/devstoreaccount1;"
+        - name: AZURE_STORAGE_CONTAINER
+          value: "ray-historyserver"
+        image: historyserver:v0.1.0
+        imagePullPolicy: IfNotPresent
+        command:
+        - historyserver
+        - --runtime-class-name=azureblob
+        - --ray-root-dir=log
+        ports:
+        - containerPort: 8080
+        resources:
+          limits:
+            cpu: "500m"

historyserver/docs/README.md

@@ -81,7 +81,8 @@ Sample configs are in the `config/` directory:
 | `raycluster.yaml` | Ray cluster with collector sidecar (S3/MinIO) |
 | `raycluster-azureblob.yaml` | Ray cluster with collector sidecar (Azure Blob) |
 | `rayjob.yaml` | Sample Ray job for testing |
-| `historyserver.yaml` | History Server deployment |
+| `historyserver.yaml` | History Server deployment (S3/MinIO) |
+| `historyserver-azureblob.yaml` | History Server deployment (Azure Blob) |
 | `service_account.yaml` | Service account for History Server |
 
 ## Additional resources

historyserver/test/e2e/historyserver_azureblob_test.go

@@ -0,0 +1,53 @@
+package e2e
+
+import (
+	"testing"
+
+	"github.com/Azure/azure-sdk-for-go/sdk/storage/azblob"
+	. "github.com/onsi/gomega"
+	corev1 "k8s.io/api/core/v1"
+
+	. "github.com/ray-project/kuberay/historyserver/test/support"
+	. "github.com/ray-project/kuberay/ray-operator/test/support"
+)
+
+func TestAzureHistoryServer(t *testing.T) {
+	azureClient := EnsureAzureBlobClient(t)
+
+	tests := []struct {
+		name     string
+		testFunc func(Test, *WithT, *corev1.Namespace, *azblob.Client)
+	}{
+		{
+			name:     "Live cluster: historyserver endpoints should be accessible",
+			testFunc: testAzureLiveClusters,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			test := With(t)
+			g := NewWithT(t)
+			namespace := test.NewTestNamespace()
+
+			tt.testFunc(test, g, namespace, azureClient)
+		})
+	}
+}
+
+func testAzureLiveClusters(test Test, g *WithT, namespace *corev1.Namespace, azureClient *azblob.Client) {
+	rayCluster := PrepareAzureHistoryServerTestEnv(test, g, namespace, azureClient)
+	ApplyRayJobAndWaitForCompletion(test, g, namespace, rayCluster)
+	ApplyAzureHistoryServer(test, g, namespace)
+	historyServerURL := GetHistoryServerURL(test, g, namespace)
+
+	clusterInfo := getClusterFromList(test, g, historyServerURL, rayCluster.Name, namespace.Name)
+	g.Expect(clusterInfo.SessionName).To(Equal(LiveSessionName), "Live cluster should have sessionName='live'")
+
+	client := CreateHTTPClientWithCookieJar(g)
+	setClusterContext(test, g, client, historyServerURL, namespace.Name, rayCluster.Name, clusterInfo.SessionName)
+	verifyHistoryServerEndpoints(test, g, client, historyServerURL)
+
+	DeleteAzureBlobContainer(test, g, azureClient)
+	LogWithTimestamp(test.T(), "Azure live clusters E2E test completed successfully")
+}

historyserver/test/support/azureblob.go

@@ -8,6 +8,7 @@ import (
 	"github.com/Azure/azure-sdk-for-go/sdk/storage/azblob"
 	. "github.com/onsi/gomega"
 	corev1 "k8s.io/api/core/v1"
+	k8serrors "k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 
 	rayv1 "github.com/ray-project/kuberay/ray-operator/apis/ray/v1"
@@ -24,7 +25,8 @@ const (
 	AzureContainerName  = "ray-historyserver"
 
 	// Azure-specific RayCluster config
-	AzureRayClusterManifestPath = "../../config/raycluster-azureblob.yaml"
+	AzureRayClusterManifestPath      = "../../config/raycluster-azureblob.yaml"
+	AzureHistoryServerManifestPath   = "../../config/historyserver-azureblob.yaml"
 )
 
 // ApplyAzurite deploys Azurite once per test namespace.
@@ -154,3 +156,59 @@ func ApplyAzureRayClusterWithCollector(test Test, g *WithT, namespace *corev1.Na
 
 	return rayCluster
 }
+
+// ApplyAzureHistoryServer deploys the HistoryServer configured for Azure Blob Storage.
+func ApplyAzureHistoryServer(test Test, g *WithT, namespace *corev1.Namespace) {
+	sa, clusterRole, clusterRoleBinding := DeserializeRBACFromYAML(test, ServiceAccountManifestPath)
+	sa.Namespace = namespace.Name
+	clusterRoleBinding.Name = fmt.Sprintf("historyserver-%s", namespace.Name)
+	clusterRoleBinding.Subjects[0].Namespace = namespace.Name
+
+	_, err := test.Client().Core().CoreV1().ServiceAccounts(namespace.Name).Create(test.Ctx(), sa, metav1.CreateOptions{})
+	g.Expect(err).NotTo(HaveOccurred())
+
+	_, err = test.Client().Core().RbacV1().ClusterRoles().Create(test.Ctx(), clusterRole, metav1.CreateOptions{})
+	if err != nil && !k8serrors.IsAlreadyExists(err) {
+		g.Expect(err).NotTo(HaveOccurred())
+	}
+
+	_, err = test.Client().Core().RbacV1().ClusterRoleBindings().Create(test.Ctx(), clusterRoleBinding, metav1.CreateOptions{})
+	g.Expect(err).NotTo(HaveOccurred())
+
+	test.T().Cleanup(func() {
+		_ = test.Client().Core().RbacV1().ClusterRoleBindings().Delete(
+			context.Background(), clusterRoleBinding.Name, metav1.DeleteOptions{})
+	})
+
+	KubectlApplyYAML(test, AzureHistoryServerManifestPath, namespace.Name)
+
+	LogWithTimestamp(test.T(), "Waiting for Azure HistoryServer to be ready")
+	g.Eventually(func(gg Gomega) {
+		pods, err := test.Client().Core().CoreV1().Pods(namespace.Name).List(
+			test.Ctx(), metav1.ListOptions{
+				LabelSelector: "app=historyserver",
+			},
+		)
+		gg.Expect(err).NotTo(HaveOccurred())
+		gg.Expect(pods.Items).NotTo(BeEmpty())
+		gg.Expect(AllPodsRunningAndReady(pods.Items)).To(BeTrue())
+	}, TestTimeoutMedium).Should(Succeed())
+	LogWithTimestamp(test.T(), "Azure HistoryServer is ready")
+}
+
+// PrepareAzureHistoryServerTestEnv prepares test environment for Azure History Server tests.
+func PrepareAzureHistoryServerTestEnv(test Test, g *WithT, namespace *corev1.Namespace, azureClient *azblob.Client) *rayv1.RayCluster {
+	rayCluster := ApplyAzureRayClusterWithCollector(test, g, namespace)
+
+	headPod, err := GetHeadPod(test, rayCluster)
+	g.Expect(err).NotTo(HaveOccurred())
+	g.Expect(headPod.Spec.Containers).To(ContainElement(
+		WithTransform(func(c corev1.Container) string { return c.Name }, Equal("collector")),
+	))
+
+	containerClient := azureClient.ServiceClient().NewContainerClient(AzureContainerName)
+	_, err = containerClient.GetProperties(context.Background(), nil)
+	g.Expect(err).NotTo(HaveOccurred())
+
+	return rayCluster
+}