GaneshPatil7517 commented on issue #17751: URL: https://github.com/apache/dolphinscheduler/issues/17751#issuecomment-3859156323
Hi @ruanwenjun, I want to be transparent with you - yes, I am using AI assistance (GitHub Copilot) to help with this contribution. I am learning the codebase and using it as a tool to understand the architecture and implement changes. I should have disclosed this upfront, and I apologize. Regarding your technical concern - you're absolutely right. I hadn't fully considered the failover scenario: The Risk: 1. handleCommand() commits transaction → WorkflowInstance persisted 2. Master crashes before onStartEvent() executes 3. New Master picks up workflow for failover 4. WorkflowExecutionGraph was never created → State inconsistency The current design where graph creation happens inside the transaction ensures atomicity - either both the WorkflowInstance AND the graph state are persisted, or neither. Given this, I have a few questions: 1.Should this issue be closed as "won't fix" since the current design is intentional for consistency? 2. Or is there an alternative approach that could reduce transaction time while maintaining consistency (perhaps caching/optimizing the graph creation rather than deferring it)? I appreciate you taking the time to explain this. It helps me understand why the current architecture is designed this 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]
