xintongsong commented on pull request #13592: URL: https://github.com/apache/flink/pull/13592#issuecomment-708932219
@tillrohrmann Thanks for the review. To answer your questions: 1. Yes, I think it is possible and makes sense. Please see my other reply for the details. 2. TBH, we do not have plan to use `allocationRequestId` at the moment. I think the priority based matching is an essential requirement for fine grained slot allocation, while the `allocationRequestId` base matching is an optimization for fine grained slot allocation with Hadoop 2.9+. Given the efforts and maintaining overhead needed for two sets of matching methods, I think we can worry about this optimization after the fine grained use cases coming true. ---------------------------------------------------------------- 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]
