dawidwys commented on pull request #18086: URL: https://github.com/apache/flink/pull/18086#issuecomment-993418208
> One more thing: shouldn't we also re-upload channel state? I.e. OperatorSubtaskState.inputChannelState and OperatorSubtaskState.resultSubpartitionState. > Or do we have a guarantee that a checkpoint can not be completed until all channel state is consumed (then I guess we need to document it somewhere). As far as I can tell: 1. channel state is never shared across checkpoints 2. in case of restore we do have a gurantee checkpoint state is consumed entirely first before we go into the `RUNNING` state. -- 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]
