fix: get pods by name in logs proxy instead of labels (#5559)

kubeshop · Jun 10, 2024 · e45318e · e45318e
1 parent c23bd57
commit e45318e
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 24 deletions.
diff --git a/pkg/executor/common.go b/pkg/executor/common.go
@@ -339,6 +339,22 @@ func IsPodFailed(pod *corev1.Pod) (err error) {
 	return
 }
 
+// GetPodByName returns job pod by name
+func GetPodByName(ctx context.Context, podsClient tcorev1.PodInterface, podName string, retryNr, retryCount int) (*corev1.Pod, error) {
+	pod, err := podsClient.Get(ctx, podName, metav1.GetOptions{})
+	if err != nil {
+		return nil, err
+	}
+	if retryNr == retryCount {
+		return nil, fmt.Errorf("retry count exceeeded, there is no active pod with given name=%s", podName)
+	}
+	if pod == nil {
+		time.Sleep(time.Duration(retryNr * 500 * int(time.Millisecond))) // increase backoff timeout
+		return GetPodByName(ctx, podsClient, podName, retryNr+1, retryCount)
+	}
+	return pod, nil
+}
+
 // GetJobPods returns job pods
 func GetJobPods(ctx context.Context, podsClient tcorev1.PodInterface, jobName string, retryNr, retryCount int) (*corev1.PodList, error) {
 	pods, err := podsClient.List(ctx, metav1.ListOptions{LabelSelector: "job-name=" + jobName})

diff --git a/pkg/logs/sidecar/proxy.go b/pkg/logs/sidecar/proxy.go
@@ -107,41 +107,38 @@ func (p *Proxy) Run(ctx context.Context) error {
 }
 
 func (p *Proxy) streamLogs(ctx context.Context, logs chan *events.Log) (err error) {
-	pods, err := executor.GetJobPods(ctx, p.podsClient, p.getPodName(), 1, 10)
+	pod, err := executor.GetPodByName(ctx, p.podsClient, p.getPodName(), 1, 10)
 	if err != nil {
 		p.handleError(err, "error getting job pods")
 		return err
 	}
 
-	for _, pod := range pods.Items {
-		l := p.log.With("namespace", pod.Namespace, "podName", pod.Name, "podStatus", pod.Status)
+	l := p.log.With("namespace", pod.Namespace, "podName", pod.Name, "podStatus", pod.Status)
 
-		switch pod.Status.Phase {
+	switch pod.Status.Phase {
 
-		case corev1.PodRunning:
-			l.Debug("streaming pod logs: immediately")
-			return p.streamLogsFromPod(pod, logs)
+	case corev1.PodRunning:
+		l.Debug("streaming pod logs: immediately")
+		return p.streamLogsFromPod(*pod, logs)
 
-		case corev1.PodFailed:
-			err := fmt.Errorf("can't get pod logs, pod failed: %s/%s", pod.Namespace, pod.Name)
-			p.handleError(err, "streaming pod logs: pod failed")
-			return err
-
-		default:
-			l.Debugw("streaming pod logs: waiting for pod to be ready")
-			testFunc := p.isPodLoggable(pod.Name)
-			if err = wait.PollUntilContextTimeout(ctx, pollInterval, podStartTimeout, true, testFunc); err != nil {
-				// try to get pod container statuses from Waiting and Terminated states
-				status := p.getPodContainerStatuses(pod)
-				p.handleError(err, "can't get pod container status after pod failure")
-				return errors.Wrap(err, status)
-			}
+	case corev1.PodFailed:
+		err := fmt.Errorf("can't get pod logs, pod failed: %s/%s", pod.Namespace, pod.Name)
+		p.handleError(err, "streaming pod logs: pod failed")
+		return err
 
-			l.Debug("streaming pod logs: pod is loggable")
-			return p.streamLogsFromPod(pod, logs)
+	default:
+		l.Debugw("streaming pod logs: waiting for pod to be ready")
+		testFunc := p.isPodLoggable(pod.Name)
+		if err = wait.PollUntilContextTimeout(ctx, pollInterval, podStartTimeout, true, testFunc); err != nil {
+			// try to get pod container statuses from Waiting and Terminated states
+			status := p.getPodContainerStatuses(*pod)
+			p.handleError(err, "can't get pod container status after pod failure")
+			return errors.Wrap(err, status)
 		}
+
+		l.Debug("streaming pod logs: pod is loggable")
+		return p.streamLogsFromPod(*pod, logs)
 	}
-	return
 }
 
 func (p *Proxy) streamLogsFromPod(pod corev1.Pod, logs chan *events.Log) (err error) {