StephanEwen commented on pull request #14199: URL: https://github.com/apache/flink/pull/14199#issuecomment-734322554
For the batch execution: Not sure over-designing the test is the right way to go here. This stuff tends to be borderline "implementation testing" not "contract testing" and gets expensive in maintenance. True, even if all slots are used, we would not necessarily get a task failure every time. But it should be enough to cause an unstable test if the batch recovery is broken. The main components of Batch recovery, also for the sources, are tested individually in a reliable way. The test here would be for the full scenario, close to the user's perspective. ---------------------------------------------------------------- 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]
