Test: Add unit test for replica validation.

This is added since we moved the validation logic.

Author: kash2104

ray-operator/controllers/ray/utils/util.go

@@ -387,15 +387,17 @@ func GenerateIdentifier(clusterName string, nodeType rayv1.RayNodeType) string {
 func GetWorkerGroupDesiredReplicas(workerGroupSpec rayv1.WorkerGroupSpec) int32 {
 	// Always adhere to min/max replicas constraints.
 	var workerReplicas int32
+	minReplicas := ptr.Deref(workerGroupSpec.MinReplicas, int32(0))
+	maxReplicas := ptr.Deref(workerGroupSpec.MaxReplicas, int32(math.MaxInt32))
 	if workerGroupSpec.Suspend != nil && *workerGroupSpec.Suspend {
 		return 0
 	}
-	if workerGroupSpec.Replicas == nil || *workerGroupSpec.Replicas < *workerGroupSpec.MinReplicas {
+	if workerGroupSpec.Replicas == nil || *workerGroupSpec.Replicas < minReplicas {
 		// Replicas is impossible to be nil as it has a default value assigned in the CRD.
 		// Add this check to make testing easier.
-		workerReplicas = *workerGroupSpec.MinReplicas
-	} else if *workerGroupSpec.Replicas > *workerGroupSpec.MaxReplicas {
-		workerReplicas = *workerGroupSpec.MaxReplicas
+		workerReplicas = minReplicas
+	} else if *workerGroupSpec.Replicas > maxReplicas {
+		workerReplicas = maxReplicas
 	} else {
 		workerReplicas = *workerGroupSpec.Replicas
 	}
@@ -419,7 +421,8 @@ func CalculateMinReplicas(cluster *rayv1.RayCluster) int32 {
 		if nodeGroup.Suspend != nil && *nodeGroup.Suspend {
 			continue
 		}
-		count += (*nodeGroup.MinReplicas * nodeGroup.NumOfHosts)
+		minReplicas := ptr.Deref(nodeGroup.MinReplicas, int32(0))
+		count += (minReplicas * nodeGroup.NumOfHosts)
 	}
 
 	return count
@@ -432,7 +435,8 @@ func CalculateMaxReplicas(cluster *rayv1.RayCluster) int32 {
 		if nodeGroup.Suspend != nil && *nodeGroup.Suspend {
 			continue
 		}
-		count += int64(*nodeGroup.MaxReplicas) * int64(nodeGroup.NumOfHosts)
+		maxReplicas := ptr.Deref(nodeGroup.MaxReplicas, int32(math.MaxInt32))
+		count += int64(maxReplicas) * int64(nodeGroup.NumOfHosts)
 	}
 
 	return SafeInt64ToInt32(count)
@@ -494,11 +498,8 @@ func CalculateMinResources(cluster *rayv1.RayCluster) corev1.ResourceList {
 	for _, nodeGroup := range cluster.Spec.WorkerGroupSpecs {
 		podResource := CalculatePodResource(nodeGroup.Template.Spec)
 		calculateReplicaResource(&podResource, nodeGroup.NumOfHosts)
-		minReplicas := int32(0)
-		if nodeGroup.MinReplicas != nil {
-			minReplicas = *nodeGroup.MinReplicas
-		}
-		for i := int32(0); i < minReplicas; i++ {
+		minReplicas := ptr.Deref(nodeGroup.MinReplicas, int32(0))
+		for range minReplicas {
 			minResourcesList = append(minResourcesList, podResource)
 		}
 	}

ray-operator/controllers/ray/utils/validation_test.go

@@ -395,7 +395,9 @@ func TestValidateRayClusterSpecEmptyContainers(t *testing.T) {
 		Template: podTemplateSpec(nil, nil),
 	}
 	workerGroupSpecWithOneContainer := rayv1.WorkerGroupSpec{
-		Template: podTemplateSpec(nil, nil),
+		Template:    podTemplateSpec(nil, nil),
+		MinReplicas: ptr.To(int32(0)),
+		MaxReplicas: ptr.To(int32(5)),
 	}
 	headGroupSpecWithNoContainers := *headGroupSpecWithOneContainer.DeepCopy()
 	headGroupSpecWithNoContainers.Template.Spec.Containers = []corev1.Container{}
@@ -459,8 +461,10 @@ func TestValidateRayClusterSpecSuspendingWorkerGroup(t *testing.T) {
 		Template: podTemplateSpec(nil, nil),
 	}
 	workerGroupSpecSuspended := rayv1.WorkerGroupSpec{
-		GroupName: "worker-group-1",
-		Template:  podTemplateSpec(nil, nil),
+		GroupName:   "worker-group-1",
+		Template:    podTemplateSpec(nil, nil),
+		MinReplicas: ptr.To(int32(0)),
+		MaxReplicas: ptr.To(int32(5)),
 	}
 	workerGroupSpecSuspended.Suspend = ptr.To(true)
 
@@ -692,8 +696,10 @@ func TestValidateRayClusterSpec_Resources(t *testing.T) {
 			},
 			WorkerGroupSpecs: []rayv1.WorkerGroupSpec{
 				{
-					GroupName: "worker-group",
-					Template:  podTemplateSpec(nil, nil),
+					GroupName:   "worker-group",
+					Template:    podTemplateSpec(nil, nil),
+					MinReplicas: ptr.To(int32(0)),
+					MaxReplicas: ptr.To(int32(5)),
 				},
 			},
 		}
@@ -773,8 +779,10 @@ func TestValidateRayClusterSpec_Labels(t *testing.T) {
 			},
 			WorkerGroupSpecs: []rayv1.WorkerGroupSpec{
 				{
-					GroupName: "worker-group",
-					Template:  podTemplateSpec(nil, nil),
+					GroupName:   "worker-group",
+					Template:    podTemplateSpec(nil, nil),
+					MinReplicas: ptr.To(int32(0)),
+					MaxReplicas: ptr.To(int32(5)),
 				},
 			},
 		}
@@ -2473,3 +2481,83 @@ func TestValidateRayClusterUpgradeOptions(t *testing.T) {
 		})
 	}
 }
+
+func TestValidateRayClusterSpec_WorkerGroupReplicaValidation(t *testing.T) {
+	createSpec := func() rayv1.RayClusterSpec {
+		return rayv1.RayClusterSpec{
+			EnableInTreeAutoscaling: ptr.To(false),
+			HeadGroupSpec: rayv1.HeadGroupSpec{
+				Template: podTemplateSpec(nil, nil),
+			},
+		}
+	}
+
+	tests := []struct {
+		name        string
+		errorMsg    string
+		spec        rayv1.RayClusterSpec
+		expectError bool
+	}{
+		{
+			name: "minReplicas greater than maxReplicas",
+			spec: func() rayv1.RayClusterSpec {
+				s := createSpec()
+				s.WorkerGroupSpecs = []rayv1.WorkerGroupSpec{
+					{
+						GroupName:   "worker-group-3",
+						Template:    podTemplateSpec(nil, nil),
+						MinReplicas: ptr.To(int32(5)),
+						MaxReplicas: ptr.To(int32(3)),
+					},
+				}
+				return s
+			}(),
+			expectError: true,
+			errorMsg:    "worker group worker-group-3 has minReplicas 5 greater than maxReplicas 3",
+		},
+		{
+			name: "replicas smaller than minReplicas when autoscaling disabled",
+			spec: func() rayv1.RayClusterSpec {
+				s := createSpec()
+				s.WorkerGroupSpecs = []rayv1.WorkerGroupSpec{
+					{
+						GroupName:   "worker-group-3",
+						Template:    podTemplateSpec(nil, nil),
+						Replicas:    ptr.To(int32(1)),
+						MinReplicas: ptr.To(int32(2)),
+						MaxReplicas: ptr.To(int32(5)),
+					},
+				}
+				return s
+			}(),
+			expectError: false,
+		},
+		{
+			name: "valid when autoscaling enabled",
+			spec: func() rayv1.RayClusterSpec {
+				s := createSpec()
+				s.EnableInTreeAutoscaling = ptr.To(true)
+				s.WorkerGroupSpecs = []rayv1.WorkerGroupSpec{
+					{
+						GroupName: "worker-group-3",
+						Template:  podTemplateSpec(nil, nil),
+					},
+				}
+				return s
+			}(),
+			expectError: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := ValidateRayClusterSpec(&tt.spec, nil)
+			if tt.expectError {
+				require.Error(t, err)
+				require.EqualError(t, err, tt.errorMsg)
+			} else {
+				require.NoError(t, err)
+			}
+		})
+	}
+}