aipatr commented on issue #19701: URL: https://github.com/apache/airflow/issues/19701#issuecomment-974113351
Use this flag: https://airflow.apache.org/docs/apache-airflow/stable/configurations-ref.html#execute-tasks-new-python-interpreter - this is slower but each task is not run in a forked process but in a separate, new interpreter. That should help. This solution doesn´t work. I set in the config the variable. Restarted Webserver and scheduler and the message continue to be: RuntimeError: Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method -- 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]
