[jira] [Created] (AIRFLOW-6810) KubernetesPodOperator pod is completed but xcom side car is stuck

["Maxence Cramet (Jira)" <ji...@apache.org>]

Maxence Cramet created AIRFLOW-6810:
---------------------------------------

             Summary: KubernetesPodOperator pod is completed but xcom side car is stuck
                 Key: AIRFLOW-6810
                 URL: https://issues.apache.org/jira/browse/AIRFLOW-6810
             Project: Apache Airflow
          Issue Type: Bug
          Components: executor-kubernetes
    Affects Versions: 1.10.6
            Reporter: Maxence Cramet
            Assignee: Daniel Imberman


We're using KubernetesPodOperator with param xcom_push=true in order to push information from our task.

From time to time the main pod completes but the side car pod is stuck.

Here's the output of the pods details:
{noformat}
kubectl describe pod my_pod
Name:               my_pod
Namespace:          default
Priority:           0
PriorityClassName:  <none>
Node:               xxx
Start Time:         Wed, 05 Feb 2020 11:12:33 +0000
Labels:             xxx
Annotations:        xxx
Status:             Running
IP:                 xxx
Containers:
  base:
    Container ID:  xxx
    Image:         xxx
    Image ID:      xxx
    Port:          <none>
    Host Port:     <none>
    Args:
      xxx
    State:          Terminated
      Reason:       Completed
      Exit Code:    0
      Started:      Wed, 05 Feb 2020 11:12:38 +0000
      Finished:     Wed, 05 Feb 2020 11:12:47 +0000
    Ready:          False
    Restart Count:  0
    Limits:
      memory:  512Mi
    Requests:
      memory:  512Mi
    Environment:
      xxx
    Mounts:
      /airflow/xcom from xcom (rw)
  airflow-xcom-sidecar:
    Container ID:  docker://83053d7d292cda9156454ac13064d72ace1e4f72738ba9b62b04ff57cb7966cc
    Image:         alpine
    Image ID:      docker-pullable://alpine@sha256:ab00606a42621fb68f2ed6ad3c88be54397f981a7b70a79db3d1172b11c4367d
    Port:          <none>
    Host Port:     <none>
    Command:
      sh
      -c
      trap "exit 0" INT; while true; do sleep 30; done;
    State:          Running
      Started:      Wed, 05 Feb 2020 11:12:40 +0000
    Ready:          True
    Restart Count:  0
    Limits:
      memory:  4Gi
    Requests:
      cpu:        1m
      memory:     2Gi
    Environment:  <none>
    Mounts:
      /airflow/xcom from xcom (rw)
      xxx
Conditions:
  Type              Status
  Initialized       True 
  Ready             False 
  ContainersReady   False 
  PodScheduled      True 
Volumes:
  xcom:
    Type:       EmptyDir (a temporary directory that shares a pod's lifetime)
    Medium:     
    SizeLimit:  <unset>
  xxx
QoS Class:       Burstable
Node-Selectors:  <none>
Tolerations:     node.kubernetes.io/not-ready:NoExecute for 300s
                 node.kubernetes.io/unreachable:NoExecute for 300s
Events:          <none>{noformat}
I don't have more information of the possible causes of that.

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)