Radeity commented on issue #12968: URL: https://github.com/apache/dolphinscheduler/issues/12968#issuecomment-1326498204
Hi @hzyangkai, thanks for your detailed design! Firstly, i strongly agree with you that all types of tasks should run as type 3 which can make DS behave more like a scheduler and without regard to how task run. Then, in my view, it's actually an ideal fault tolerance strategy and better than current way. However, i have a question that how can you exit submit process after fetching appId immediately and how to separate `submitApplication` and `monitorApplication` ? In addition, current way to fetch appId is to parse from log file when finish executing or by `GET_APP_ID_REQUEST` request. Although i provide an aop way to collect appId(https://github.com/apache/dolphinscheduler/pull/12197), it’s not suitable for all yarn tasks such as submitting yarn job on remote host by client mode like Beeline which you have to consider, either. If you can separate two steps mentioned above, you can collect appId more efficiently than current way. -- 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]
