fix(logservice): eliminate RunJob false-success and harden CreatePVC

- RunJob: timeout and empty-pod-list paths used errors.Wrapf(err, ...)
  which returns nil when err is nil, causing RunJob to return a nil
  error (false success) when job polling timed out with the last GetJob
  succeeding. This made LogService bootstrap falsely report "succeeded"
  while the job was still failing. Use errors.Errorf for the timeout
  path and split the pod-list condition so nil err is not wrapped.
- CreatePVC: snapshot StoreStorage/LogStorage into locals right after
  the nil checks and use them in PVC construction, defending against an
  intermittent nil-deref observed at the LogStorage.Size use despite the
  pre-existing nil checks.

Author: Junkrat77

internal/resource/oblogservicenode/oblogservicenode_task.go

@@ -104,6 +104,13 @@ func CreatePVC(m *OBLogServiceNodeManager) tasktypes.TaskError {
 		return errors.New("storage.logStorage is required but was nil")
 	}
 
+	// Snapshot the storage fields into locals after the nil checks so that any
+	// hypothetical concurrent mutation of m.Resource.Spec.Storage between the
+	// checks and the PVC construction cannot turn a non-nil check into a nil
+	// deref (observed as an intermittent panic at the LogStorage.Size use).
+	storeStorage := m.Resource.Spec.Storage.StoreStorage
+	logStorage := m.Resource.Spec.Storage.LogStorage
+
 	podName := m.Resource.Name
 	storePvcName := fmt.Sprintf("%s-%s", podName, oceanbaseconst.LogServiceStoreVolumeSuffix)
 	logPvcName := fmt.Sprintf("%s-%s", podName, oceanbaseconst.LogServiceLogVolumeSuffix)
@@ -127,10 +134,10 @@ func CreatePVC(m *OBLogServiceNodeManager) tasktypes.TaskError {
 			AccessModes: []corev1.PersistentVolumeAccessMode{corev1.ReadWriteOnce},
 			Resources: corev1.VolumeResourceRequirements{
 				Requests: corev1.ResourceList{
-					"storage": m.Resource.Spec.Storage.StoreStorage.Size,
+					"storage": storeStorage.Size,
 				},
 			},
-			StorageClassName: &m.Resource.Spec.Storage.StoreStorage.StorageClass,
+			StorageClassName: &storeStorage.StorageClass,
 		},
 	}
 	if err := m.Client.Create(m.Ctx, storePvc); err != nil && !kubeerrors.IsAlreadyExists(err) {
@@ -147,10 +154,10 @@ func CreatePVC(m *OBLogServiceNodeManager) tasktypes.TaskError {
 			AccessModes: []corev1.PersistentVolumeAccessMode{corev1.ReadWriteOnce},
 			Resources: corev1.VolumeResourceRequirements{
 				Requests: corev1.ResourceList{
-					"storage": m.Resource.Spec.Storage.LogStorage.Size,
+					"storage": logStorage.Size,
 				},
 			},
-			StorageClassName: &m.Resource.Spec.Storage.LogStorage.StorageClass,
+			StorageClassName: &logStorage.StorageClass,
 		},
 	}
 	if err := m.Client.Create(m.Ctx, logPvc); err != nil && !kubeerrors.IsAlreadyExists(err) {

internal/resource/utils/jobs.go

@@ -120,14 +120,20 @@ func RunJob(ctx context.Context, c client.Client, logger *logr.Logger, namespace
 		time.Sleep(time.Second * time.Duration(obcfg.GetConfig().Time.CheckJobInterval))
 	}
 	if !finished {
-		return "", int32(cmdconst.ExitCodeNotExecuted), errors.Wrapf(err, "Run job %s timeout", fullJobName)
+		// NOTE: do not use errors.Wrapf here — when the last GetJob succeeded
+		// (job still running), err is nil and Wrapf(nil, ...) returns nil,
+		// which would make RunJob falsely report success.
+		return "", int32(cmdconst.ExitCodeNotExecuted), errors.Errorf("Run job %s timeout", fullJobName)
 	}
 	clientSet := k8sclient.GetClient()
 	podList, err := clientSet.ClientSet.CoreV1().Pods(namespace).List(ctx, metav1.ListOptions{
 		LabelSelector: fmt.Sprintf("job-name=%s", fullJobName),
 	})
 	if err != nil || len(podList.Items) == 0 {
-		return "", int32(cmdconst.ExitCodeNotExecuted), errors.Wrapf(err, "failed to get pods of job %s", jobName)
+		if err != nil {
+			return "", int32(cmdconst.ExitCodeNotExecuted), errors.Wrapf(err, "failed to get pods of job %s", jobName)
+		}
+		return "", int32(cmdconst.ExitCodeNotExecuted), errors.Errorf("failed to get pods of job %s: no pods found", jobName)
 	}
 	var outputBuffer bytes.Buffer
 	podLogOpts := corev1.PodLogOptions{}