Re: Ambiguous behavior of Flink on Job cancellation with checkpoint configured

That makes sense. Thanks for the clarification. -- Sent from: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/

Re: Ambiguous behavior of Flink on Job cancellation with checkpoint configured

Hi, That’s a good observation! And it is indeed the expected behaviour. There are two parts to understanding this: * "retain checkpoints” tells Flink to retain any checkpoints that it stores when a job is shut down * for recovery purposes (all checkpointing purposes, really) a savepoint count