[jira] [Created] (AIRFLOW-2922) Potential deal-lock bug in CeleryExecutor()

["Xiaodong DENG (JIRA)" <ji...@apache.org>]

Xiaodong DENG created AIRFLOW-2922:
--------------------------------------

             Summary: Potential deal-lock bug in CeleryExecutor()
                 Key: AIRFLOW-2922
                 URL: https://issues.apache.org/jira/browse/AIRFLOW-2922
             Project: Apache Airflow
          Issue Type: Improvement
          Components: executor
            Reporter: Xiaodong DENG
            Assignee: Xiaodong DENG


Code: [https://github.com/apache/incubator-airflow/blob/404be4b021e803239f48e4dd5705759cc77aeb92/airflow/executors/celery_executor.py] 

Celery Task states normally change in ways like “PENDING -> STARTED -> SUCCESS/FAILURE” ([http://docs.celeryproject.org/en/latest/reference/celery.states.html]).

In lines 107 and 108, it’s `task.state` rather than `state`, i.e. it will reflect the latest real-time state of the Celery task.

Let’s imagine: task state becomes “STARTED” (initial state is “PENDING”), then the if-elif-else block will be triggered. It’s possible the Celery task state becomes “SUCCESS” when whichever line between 94-105 is running. At line 108, the latest state of the task in _*self.last_state*_ will be changed to “SUCCESS” rather than “STARTED” because it’s referring to `task.state` rather than variable `state`.

Then this task will be dead-locked because the if-elif-else block will never be triggered for it.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)