smjn opened a new pull request, #23112: URL: https://github.com/apache/kafka/pull/23112
Similar to https://github.com/apache/kafka/pull/23074, the writeState handler in the ShareCoordinator, directly writes the `deliveryCompleteCount` into the state. This might result in deliveryCompleteCount regressing momentarily if an out or order request with a lower delivery complete count than the actual value updates the state. This PR remedies the issue by leveraging the start offset - based on the start offset value in the incoming request, we should choose the delivery complete count value: - incoming(startOffset) > current(startOffset) - take incoming delivery complete count - incoming(startOffset) == current(startOffset) - take Max(incoming delivery complete count, current delivery complete count) - incoming(startOffset) < current(startOffset) - take current delivery count We cannot compare the delivery complete counts directly as they are non non-monotonic sequences. -- 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]
