Jorricks removed a comment on issue #16023: URL: https://github.com/apache/airflow/issues/16023#issuecomment-876729124
I did some checks. I think it might be related to the fact that when running through the webserver, we never set the `external_executor_id` value. In a Scheduler run, this is set after a the task is queued, at which point the scheduler will read the event_buffer and set the `external_executor_id` based on that. However on the webserver side, this loop isn't present and `external_executor_id` isn't set, meaning when another scheduler picks it up, it dies. We should **not** see this behaviour for tasks that are first started by the Scheduler and then (re-)triggered manually. Code fragment of where we set it when tasks is started by the Scheduler: https://github.com/apache/airflow/blob/db6acd9e8a91e0eca9e12cace72edc57b2667d25/airflow/jobs/scheduler_job.py#L596-L599 To verify my theory, it would help if you could provide DEBUG logs of the scheduler and the webserver while reproducing the issue. Could you please be so kind to provide these when you have the time? -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
