raj-manvar commented on issue #9610:
URL: https://github.com/apache/airflow/issues/9610#issuecomment-689941118


   Few new findings:
   * was able to reproduce this issue for task running more than a minute by 
setting `visibility_timeout = 60` in [celery_broker_transport_options] in 
airflow.cfg file. 
   * Therefore, this is happening because Celery expects the task to complete 
within an hour and if not assigns another worker for the task, during this 
transition, worker uploads the logs with "Task is not able to run" to S3
   * Can see another worker getting same task from logs `Received task: 
airflow.executors.celery_executor.execute_command[b40cacbb-9dd3-4681-8454-0e1df2dbc910]`
 with same id seconding that Celery is assigning this task to another worker.
   
   * Modifying "visibility_timeout = 86400 # 1day" in airflow.cfg doesn't 
resolve this issue and logs in UI are corrupted after an hour
   * Even tried "visibility_timeout = 7200 # 2 hours" in airflow.cfg but can 
still see this issue after an hour. 
   * Seems the issue is similar to 
https://github.com/celery/celery/issues/5935, but according to this it should 
be resolved in Celery version 4.4.5 but, we still see the same issue even 
though Airflow 1.10.10 uses Celery version 4.4.6 
   
   ( CC: @chrismclennon @dimberman  ) 


----------------------------------------------------------------
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