xloya opened a new issue #4208: URL: https://github.com/apache/incubator-dolphinscheduler/issues/4208
**Describe the question** Now DS's worker fault tolerance strategy will kill the running yarn app and resubmit when worker failover,but there are a lot of scenarios that users don't want to resubmit their yarn apps and keep them running.And if DS wants to support real-time task in the future,this is also not an appropriate strategy. **What are the current deficiencies and the benefits of improvement** Can we consider changing the fault tolerance strategy of worker to the app that is already running and let it continue running, and then transfer the yarn app to other workers to continue listening？ ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
