Airflow kubernetes吊舱操作员重试不工作

我正在Kubernetes群集上以CeleryExecutor模式运行气流1.10.12。此DAG将在同一Kubernetes群集上但在不同名称空间中创建一个Pod

如果运行良好，pod将成功运行spark作业。但如果失败，应该重试。通过重试，失败的pod应该会收到一个新标签metadata.labels.ready_checked，第一次重试时为“True”，第二次重试时应启动一个新的pod。但是，新标签未按预期标记

错误消息的快照

Traceback (most recent call last):
  File "/usr/local/lib/python3.6/site-packages/airflow/contrib/operators/kubernetes_pod_operator.py", line 282, in execute
    final_state, result = self.handle_pod_overlap(labels, try_numbers_match, launcher, pod_list)
  File "/usr/local/lib/python3.6/site-packages/airflow/contrib/operators/kubernetes_pod_operator.py", line 312, in handle_pod_overlap
    final_state, result = self.monitor_launched_pod(launcher, pod_list.items[0])
  File "/usr/local/lib/python3.6/site-packages/airflow/contrib/operators/kubernetes_pod_operator.py", line 432, in monitor_launched_pod
    'Pod returned a failure: {state}'.format(state=final_state)
airflow.exceptions.AirflowException: Pod returned a failure: failed

---

失败的旧吊舱还活着吗？我试图通过在旧吊舱终止后清除它来重试气流任务。@HenryCarter您能看到以下问题是否与您的案例相关吗？失败的旧吊舱还活着吗？我试图在旧吊舱终止后清除它来重试气流任务。@HenryCarter您能看到以下问题是否存在吗和你的案子有关吗？