Arunodoy18 commented on issue #60943: URL: https://github.com/apache/airflow/issues/60943#issuecomment-3796049700
This looks like a concurrency issue rather than a stale cache problem. When multiple KubernetesPodOperator tasks start in parallel on the same Celery worker, they each invoke aws eks get-token, and the AWS CLI attempts to create ~/.aws/cli/cache concurrently. Since the AWS CLI does not guard directory creation, a race condition can occur, resulting in FileExistsError before pod creation. The failure happens during Kubernetes API authentication, which explains why it surfaces as a 403. This seems reproducible under parallel KPO execution on long-lived workers. -- 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]
