morhidi commented on PR #557: URL: https://github.com/apache/flink-kubernetes-operator/pull/557#issuecomment-1490779906
> > @gyfora could you please shed some light on the benefits of this feature? > > If the checkpoint interval is high (lets say 1 hour) it can happen that we trigger a last state upgrade close to the end of the interval simply falling back to the last checkpoint. This would mean a very large backlog to process, probably unintentionally. > > This feature would allow us to automatically take a savepoint instead (or in Flink 1.17+ trigger a checkpoint) > > @gyfora could you please shed some light on the benefits of this feature? > > If the checkpoint interval is high (lets say 1 hour) it can happen that we trigger a last state upgrade close to the end of the interval simply falling back to the last checkpoint. This would mean a very large backlog to process, probably unintentionally. > > This feature would allow us to automatically take a savepoint instead (or in Flink 1.17+ trigger a checkpoint) Thanks for the clarification. I guess users are able to tune their checkpoint interval to avoid scenario we're trying to address with this feature. I like the idea of checkpoint upgrade mode though. -- 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]
