haseebmalik18 commented on PR #62823: URL: https://github.com/apache/airflow/pull/62823#issuecomment-4000262234
> To be honest, I don't like this approach personally. Can we find a way to make sure the supervisor can get a new connection from the connection Pool? Pool_pre_ping is already enabled but it only checks connections at checkout. Error 2013 means the connection dies mid-query so pre-ping can't catch it. SQLAlchemy does invalidate the dead connection in the pool so the next one is fresh, but the current operation still fails. If I'm not mistaken, I don't think there is a pool config that can recover a failed mid-query operation, it has to be re-executed which is what the retry does here. If you have a approach in mind though I'd love to hear it, happy to collaborate on this together. -- 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]
