JunRuiLee commented on PR #24173: URL: https://github.com/apache/flink/pull/24173#issuecomment-1907724157
> @JunRuiLee Thanks for joining the disscussion . > > > the PR skips the creation of the checkpoint coordinator when disabling the checkpoint scheduling > > Yes, actually I think that's also why we made this change. Because Flink does not require the creation of components such as CheckpointCoordinator or CheckpointIDCounter in many scenarios, such as OLAP or Flink Batch. In these scenarios, I believe it is reasonable not to create a CheckpointCoordinator. > > > For instance, could this also prevent triggering savepoints? > > Yes. If disabling Checkpoint, the savepoint cannot be triggered after Change. But I think this issue is indeed worth discussing. In my opinion, if the user disable Checkpoint, does it mean that the user can accept state loss. I can't imagine a scenario where users don't need Checkpoint but need Savepoint. So i thinks it's fine here or we need notice the user savepoint can only be triggered while Checkpoint is enabled . WDYT In my opinion, disabling checkpoints does not necessarily equate to disabling savepoints or other functionalities of the Checkpoint Coordinator, such as restoring from state. Some users may prefer to avoid the automatic triggering of checkpoints, yet still retain the option to initiate savepoints manually when needed. -- 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]
