zhijiangW commented on issue #7186: [FLINK-10941] Keep slots which contain unconsumed result partitions URL: https://github.com/apache/flink/pull/7186#issuecomment-468538365 @azagrebin , you pointed out a very good question which I have not thought before. :) Currently app/flink does not provide the explicit ack mechanisms to confirm the received message in network. We only rely on the TCP internal ack mechanisms. For data message (`BufferResponse`) transfer, we add the sequence number mechanism on sender&receiver side to verify there are no messages lost in network, otherwise it would cause failover. For other protocol messages (`PartitionRequest`, `CancelRequest`, `CloseRequest`, etc) we have not the other mechanisms to confirm that if the network condition is not reliable. But for `CloseRequest`, I think it would not cause problems currently, because the consumer would close the connection after all, then the producer would be aware of inactive channel to release all the resources. But your reminder makes me thought of potential problems for other messages.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services