You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@airflow.apache.org by GitBox <gi...@apache.org> on 2021/01/12 10:10:14 UTC

[GitHub] [airflow] Marzocchi666 commented on issue #10964: No response from gunicorn master within 120 seconds After Changing Worker Class

Marzocchi666 commented on issue #10964:
URL: https://github.com/apache/airflow/issues/10964#issuecomment-758550957


   > I had the same issue and noticed in the webserver debug logs that after spinning up the workers they went immediately into refreshing which seems to cause a crash eg:
   > 
   > ```
   > [3 / 4] some workers are starting up, waiting...
   > [4 / 4] sleeping for 120s starting doing a refresh...
   > [4 / 4] doing a refresh of 1 workers
   > [4 / 5] some workers are starting up, waiting...
   > ...
   > [5 / 5] killing 1 workers
   > Handling signal: ttou
   > Worker exiting (pid: 164)
   > No response from gunicorn master within 120 seconds
   > ```
   > 
   > I was able to work around this by setting `worker_refresh_batch_size` and `worker_refresh_interval` to 0.
   > 
   > This was using gevent as well on airflow 1.10.10
   
   interesting, because if you switch worker_refresh_batch_size to 0 means disable the refresh.


----------------------------------------------------------------
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:
users@infra.apache.org