FloChehab opened a new issue #10541:
URL: https://github.com/apache/airflow/issues/10541


   **Apache Airflow version**: 1.10.12 rc4
   
   **Kubernetes version (if you are using kubernetes)** (use `kubectl 
version`):  v1.16.11-gke.5
   
   **Environment**:
   
   - **Cloud provider or hardware configuration**: /
   - **OS** (e.g. from /etc/os-release): /
   - **Kernel** (e.g. `uname -a`): /
   - **Install tools**: /
   - **Others**: 
`apache/airflow@sha256:6de1374274f26836c98bbe9f8c065215491f8f5bd48bedc155765dec9b883144`
   
   **What happened**:
   
   This issue is a followup to discussions on 
https://github.com/apache/airflow/pull/10230#issuecomment-679274286 .
   
   Let's take this dag:
   
   ```python
   from airflow.contrib.operators.kubernetes_pod_operator import 
KubernetesPodOperator
   from airflow.kubernetes.secret import Secret
   from airflow.models import DAG
   from airflow.utils.dates import days_ago
   
   
   default_args = {
       'owner': 'Airflow',
       'start_date': days_ago(2),
       'retries': 3
   }
   
   with DAG(
       dag_id='bug_kuberntes_pod_operator',
       default_args=default_args,
       schedule_interval=None
   ) as dag:
       k = KubernetesPodOperator(
           namespace='airflow',
           image="ubuntu:16.04",
           cmds=["bash", "-cx"],
           arguments=["sleep 100"],
           name="airflow-test-pod",
           task_id="task",
           get_logs=True,
           is_delete_operator_pod=True,
       )
   ```
   
   If you:
   1. Trigger the dag,
   2. Wait for the task to be up and running on kubernetes,
   3. Kill everything related to airflow (except the task running on 
kubernetes),
   4. Wait for the task to complete on Kubernetes,
   5. Restart airflow.
   
   The the task would be marked as `up_for_retry` and would be stuck in this 
state until another scheduler restart.
   
   **What you expected to happen**:
   
   The task to be marked as success on the first scheduler restart or not stuck 
in `up_for_retry` state.
   
   **How to reproduce it**:
   
   * Use the dag above,
   * Tested with LocalExecutor and CeleryExecutor (on keda) ; both with helm 
chart from master. With no major changes except setting the timezone to 
Europe/Paris.
   
   
   **Anything else we need to know**:
   
   * The issue seems to appear every time,
   * Scheduler logs can be found here: 
https://github.com/apache/airflow/pull/10230#issuecomment-679304807 & 
https://github.com/apache/airflow/pull/10230#issuecomment-679314891


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to