Merge pull request #289 from lminzhw/add_unknown_pod

count unknown pod as [Unknown] in job status

pkg/apis/batch/v1alpha1/job.go

@@ -285,15 +285,19 @@ type JobStatus struct {
 	// +optional
 	Terminating int32 `json:"terminating,omitempty" protobuf:"bytes,7,opt,name=terminating"`
 
+	// The number of pods which reached phase Unknown.
+	// +optional
+	Unknown int32 `json:"unknown,omitempty" protobuf:"bytes,8,opt,name=unknown"`
+
 	//Current version of job
-	Version int32 `json:"version,omitempty" protobuf:"bytes,8,opt,name=version"`
+	Version int32 `json:"version,omitempty" protobuf:"bytes,9,opt,name=version"`
 
 	// The number of Job retries.
 	// +optional
-	RetryCount int32 `json:"retryCount,omitempty" protobuf:"bytes,9,opt,name=retryCount"`
+	RetryCount int32 `json:"retryCount,omitempty" protobuf:"bytes,10,opt,name=retryCount"`
 
 	// The resources that controlled by this job, e.g. Service, ConfigMap
-	ControlledResources map[string]string `json:"controlledResources,omitempty" protobuf:"bytes,8,opt,name=controlledResources"`
+	ControlledResources map[string]string `json:"controlledResources,omitempty" protobuf:"bytes,11,opt,name=controlledResources"`
 }
 
 // +k8s:deepcopy-gen:interfaces=k8s.io/apimachinery/pkg/runtime.Object

pkg/cli/job/list.go

@@ -65,6 +65,8 @@ const (
 	Version string = "Version"
 	// Failed  failed
 	Failed string = "Failed"
+	// Unknown pod
+	Unknown string = "Unknown"
 	// RetryCount retry count
 	RetryCount string = "RetryCount"
 	// JobType  job type
@@ -110,8 +112,8 @@ func ListJobs() error {
 // PrintJobs prints all jobs details
 func PrintJobs(jobs *v1alpha1.JobList, writer io.Writer) {
 	maxNameLen := getMaxNameLen(jobs)
-	_, err := fmt.Fprintf(writer, fmt.Sprintf("%%-%ds%%-25s%%-12s%%-12s%%-12s%%-6s%%-10s%%-10s%%-12s%%-10s%%-12s\n", maxNameLen),
-		Name, Creation, Phase, JobType, Replicas, Min, Pending, Running, Succeeded, Failed, RetryCount)
+	_, err := fmt.Fprintf(writer, fmt.Sprintf("%%-%ds%%-25s%%-12s%%-12s%%-12s%%-6s%%-10s%%-10s%%-12s%%-10s%%-12s%%-10s\n", maxNameLen),
+		Name, Creation, Phase, JobType, Replicas, Min, Pending, Running, Succeeded, Failed, Unknown, RetryCount)
 	if err != nil {
 		fmt.Printf("Failed to print list command result: %s.\n", err)
 	}
@@ -131,9 +133,9 @@ func PrintJobs(jobs *v1alpha1.JobList, writer io.Writer) {
 		if jobType == "" {
 			jobType = "Batch"
 		}
-		_, err = fmt.Fprintf(writer, fmt.Sprintf("%%-%ds%%-25s%%-12s%%-12s%%-12d%%-6d%%-10d%%-10d%%-12d%%-10d%%-12d\n", maxNameLen),
+		_, err = fmt.Fprintf(writer, fmt.Sprintf("%%-%ds%%-25s%%-12s%%-12s%%-12d%%-6d%%-10d%%-10d%%-12d%%-10d%%-12d%%-10d\n", maxNameLen),
 			job.Name, job.CreationTimestamp.Format("2006-01-02 15:04:05"), job.Status.State.Phase, jobType, replicas,
-			job.Status.MinAvailable, job.Status.Pending, job.Status.Running, job.Status.Succeeded, job.Status.Failed, job.Status.RetryCount)
+			job.Status.MinAvailable, job.Status.Pending, job.Status.Running, job.Status.Succeeded, job.Status.Failed, job.Status.Unknown, job.Status.RetryCount)
 		if err != nil {
 			fmt.Printf("Failed to print list command result: %s.\n", err)
 		}

pkg/cli/job/view.go

@@ -174,6 +174,9 @@ func PrintJobInfo(job *v1alpha1.Job, writer io.Writer) {
 	if job.Status.Terminating > 0 {
 		WriteLine(writer, Level1, "Terminating:  \t%d\n", job.Status.Terminating)
 	}
+	if job.Status.Unknown > 0 {
+		WriteLine(writer, Level1, "Unknown:      \t%d\n", job.Status.Unknown)
+	}
 	if job.Status.RetryCount > 0 {
 		WriteLine(writer, Level1, "RetryCount:   \t%d\n", job.Status.RetryCount)
 	}

pkg/cli/job/view_test.go

@@ -76,6 +76,7 @@ func TestViewJob(t *testing.T) {
 			Pending:      3,
 			Running:      1,
 			Failed:       2,
+			Unknown:      10,
 			Terminating:  4,
 			RetryCount:   5,
 			MinAvailable: 6,

pkg/controllers/job/job_controller_actions.go

@@ -48,7 +48,7 @@ func (cc *Controller) killJob(jobInfo *apis.JobInfo, podRetainPhase state.PhaseM
 		return nil
 	}
 
-	var pending, running, terminating, succeeded, failed int32
+	var pending, running, terminating, succeeded, failed, unknown int32
 
 	var errs []error
 	var total int
@@ -76,16 +76,7 @@ func (cc *Controller) killJob(jobInfo *apis.JobInfo, podRetainPhase state.PhaseM
 				cc.resyncTask(pod)
 			}
 
-			switch pod.Status.Phase {
-			case v1.PodRunning:
-				running++
-			case v1.PodPending:
-				pending++
-			case v1.PodSucceeded:
-				succeeded++
-			case v1.PodFailed:
-				failed++
-			}
+			classifyAndAddUpPodBaseOnPhase(pod, &pending, &running, &succeeded, &failed, &unknown)
 		}
 	}
 
@@ -108,6 +99,7 @@ func (cc *Controller) killJob(jobInfo *apis.JobInfo, podRetainPhase state.PhaseM
 		Succeeded:    succeeded,
 		Failed:       failed,
 		Terminating:  terminating,
+		Unknown:      unknown,
 		Version:      job.Status.Version,
 		MinAvailable: int32(job.Spec.MinAvailable),
 		RetryCount:   job.Status.RetryCount,
@@ -217,7 +209,7 @@ func (cc *Controller) syncJob(jobInfo *apis.JobInfo, updateStatus state.UpdateSt
 		return nil
 	}
 
-	var running, pending, terminating, succeeded, failed int32
+	var running, pending, terminating, succeeded, failed, unknown int32
 
 	var podToCreate []*v1.Pod
 	var podToDelete []*v1.Pod
@@ -250,16 +242,7 @@ func (cc *Controller) syncJob(jobInfo *apis.JobInfo, updateStatus state.UpdateSt
 					continue
 				}
 
-				switch pod.Status.Phase {
-				case v1.PodPending:
-					pending++
-				case v1.PodRunning:
-					running++
-				case v1.PodSucceeded:
-					succeeded++
-				case v1.PodFailed:
-					failed++
-				}
+				classifyAndAddUpPodBaseOnPhase(pod, &pending, &running, &succeeded, &failed, &unknown)
 			}
 		}
 
@@ -273,7 +256,7 @@ func (cc *Controller) syncJob(jobInfo *apis.JobInfo, updateStatus state.UpdateSt
 	for _, pod := range podToCreate {
 		go func(pod *v1.Pod) {
 			defer waitCreationGroup.Done()
-			_, err := cc.kubeClients.CoreV1().Pods(pod.Namespace).Create(pod)
+			newPod, err := cc.kubeClients.CoreV1().Pods(pod.Namespace).Create(pod)
 			if err != nil && !apierrors.IsAlreadyExists(err) {
 				// Failed to create Pod, waitCreationGroup a moment and then create it again
 				// This is to ensure all podsMap under the same Job created
@@ -286,8 +269,7 @@ func (cc *Controller) syncJob(jobInfo *apis.JobInfo, updateStatus state.UpdateSt
 					cc.resyncTask(pod)
 				}
 
-				// TODO: maybe not pending status, maybe unknown.
-				pending++
+				classifyAndAddUpPodBaseOnPhase(newPod, &pending, &running, &succeeded, &failed, &unknown)
 				glog.V(3).Infof("Created Task <%s> of Job <%s/%s>",
 					pod.Name, job.Namespace, job.Name)
 			}
@@ -340,6 +322,7 @@ func (cc *Controller) syncJob(jobInfo *apis.JobInfo, updateStatus state.UpdateSt
 		Succeeded:           succeeded,
 		Failed:              failed,
 		Terminating:         terminating,
+		Unknown:             unknown,
 		Version:             job.Status.Version,
 		MinAvailable:        int32(job.Spec.MinAvailable),
 		ControlledResources: job.Status.ControlledResources,
@@ -559,3 +542,19 @@ func (cc *Controller) initJobStatus(job *vkv1.Job) (*vkv1.Job, error) {
 
 	return newJob, nil
 }
+
+func classifyAndAddUpPodBaseOnPhase(pod *v1.Pod, pending, running, succeeded, failed, unknown *int32) {
+	switch pod.Status.Phase {
+	case v1.PodPending:
+		*pending++
+	case v1.PodRunning:
+		*running++
+	case v1.PodSucceeded:
+		*succeeded++
+	case v1.PodFailed:
+		*failed++
+	default:
+		*unknown++
+	}
+	return
+}