CloudSen commented on issue #16369: URL: https://github.com/apache/dolphinscheduler/issues/16369#issuecomment-2295682973
> Right now the serial wait implementation is really unstable, there are a lot of case will cause it doesn't work well. e.g. > > 1. Concurrent trigger will cause multiple workflow instance running which should in serial wait. > 2. Notify failed might cause the origin workflow instance cannot finish. > 3. The workflow should deal with the notify logic, this make the workflow instance state transition more complex. > > It's better to refactor this, use a global SerialWaitCoordinator to notify the serial wait workflow instance, the origin workflow instance don't need to care whether it need to do notification. Agree with your opinion, I have thought about a coordinator similar to SerialWaitCoordinator instead of notifying tasks themselves. - Is there a refactoring plan for this? - How should this bug in 3.1.x be fixed? As I said, currently I can temporarily solve most problems by re-sending the RECOVER_SERIAL_WAIT command -- 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]
