Fixed PR comments

helm-chart/kuberay-operator/crds/ray.io_rayjobs.yaml

@@ -11745,7 +11745,7 @@ spec:
                 type: boolean
               suspend:
                 description: suspend specifies whether the RayJob controller should
-                  create a RayCluster instance
+                  create a RayCluster instance If a job is appl
                 type: boolean
               ttlSecondsAfterFinished:
                 description: TTLSecondsAfterFinished is the TTL to clean up RayCluster.

ray-operator/config/crd/bases/ray.io_rayjobs.yaml

@@ -11745,7 +11745,7 @@ spec:
                 type: boolean
               suspend:
                 description: suspend specifies whether the RayJob controller should
-                  create a RayCluster instance
+                  create a RayCluster instance If a job is appl
                 type: boolean
               ttlSecondsAfterFinished:
                 description: TTLSecondsAfterFinished is the TTL to clean up RayCluster.

ray-operator/controllers/ray/rayjob_controller.go

@@ -243,10 +243,10 @@ func (r *RayJobReconciler) Reconcile(ctx context.Context, request ctrl.Request)
 					return ctrl.Result{RequeueAfter: RayJobDefaultRequeueDuration}, err
 				}
 			}
-			err = r.updateState(ctx, rayJobInstance, jobInfo, rayv1alpha1.JobStatusStopped, rayJobInstance.Status.JobDeploymentStatus, nil)
-			if err != nil {
-				return ctrl.Result{RequeueAfter: RayJobDefaultRequeueDuration}, err
+			if info.JobStatus != rayv1alpha1.JobStatusStopped {
+				return ctrl.Result{RequeueAfter: RayJobDefaultRequeueDuration}, nil
 			}
+
 			_, err = r.deleteCluster(ctx, rayJobInstance)
 			if err != nil && !errors.IsNotFound(err) {
 				return ctrl.Result{RequeueAfter: RayJobDefaultRequeueDuration}, nil

ray-operator/controllers/ray/rayjob_controller_suspended_test.go

@@ -20,6 +20,7 @@ import (
 	"fmt"
 	"time"
 
+	"k8s.io/apimachinery/pkg/api/errors"
 	"k8s.io/apimachinery/pkg/api/resource"
 
 	"github.com/ray-project/kuberay/ray-operator/controllers/ray/common"
@@ -51,7 +52,7 @@ var _ = Context("Inside the default namespace", func() {
 			Suspend:    true,
 			Entrypoint: "sleep 999",
 			RayClusterSpec: &rayiov1alpha1.RayClusterSpec{
-				RayVersion: "1.12.1",
+				RayVersion: "2.4.0",
 				HeadGroupSpec: rayiov1alpha1.HeadGroupSpec{
 					ServiceType: corev1.ServiceTypeClusterIP,
 					Replicas:    pointer.Int32(1),
@@ -204,8 +205,11 @@ var _ = Context("Inside the default namespace", func() {
 			// However the actual cluster instance and underlying resources should not be created while suspend == true
 			Eventually(
 				// k8sClient client throws error if resource not found
-				getResourceFunc(ctx, client.ObjectKey{Name: mySuspendedRayJob.Status.RayClusterName, Namespace: "default"}, mySuspendedRayCluster),
-				time.Second*10, time.Millisecond*500).Should(Not(BeNil()))
+				func() bool {
+					err := getResourceFunc(ctx, client.ObjectKey{Name: mySuspendedRayJob.Status.RayClusterName, Namespace: "default"}, mySuspendedRayCluster)()
+					return errors.IsNotFound(err)
+				},
+				time.Second*10, time.Millisecond*500).Should(BeTrue())
 		})
 
 		It("should unsuspend a rayjob object", func() {

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

@@ -90,7 +90,7 @@ var _ = Describe("RayFrameworkGenerator", func() {
 		Expect(rayJobInfo.Entrypoint).To(Equal(rayJob.Spec.Entrypoint))
 		Expect(rayJobInfo.JobStatus).To(Equal(rayv1alpha1.JobStatusRunning))
 
-		_, err = rayDashboardClient.GetJobInfo(errorJobId)
+		_, err = rayDashboardClient.GetJobInfo(context.TODO(), errorJobId)
 		Expect(err).NotTo(BeNil())
 		Expect(err.Error()).To(ContainSubstring("GetJobInfo fail"))
 		Expect(err.Error()).To(ContainSubstring("Ray misbehaved"))