Zakelly opened a new pull request, #25816:
URL: https://github.com/apache/flink/pull/25816
## What is the purpose of the change
We have introduced the `EpochManager` to handle the watermark during async
state processing. The inputs are divided into several epochs where a action is
attached and triggered after all the inputs finished their async processing.
However, the action itself may trigger more async procedures which blocks some
synchronous action. Thus this PR introduces two types of action, one is for
triggered right after all the inputs finished, while another is for being
triggered after all the async processing including the ones from the action
finished.
This is a complete solution & abstraction of watermark handling. I'll revert
some change from FLINK-36875, removing the unnecessary use of
`CompletableFuture`.
## Brief change log
- Remove `CompletableFuture` in signatures of `advanceWatermark`.
- Introduce `triggerAction` and `finalAction` for each `Epoch`
- Proper handle the state machine of `Epoch`
- Support inheritance of `Epoch` for `asyncProcessWithKey`, meaning that
the context during `asyncProcessWithKey` will share the same `epoch` from
invoker's context.
## Verifying this change
This change added two tests
`AbstractAsyncStateStreamOperatorTest#testWatermark` and
`EpochManagerTest#testTwoAction`.
## Does this pull request potentially affect one of the following parts:
- Dependencies (does it add or upgrade a dependency): no
- The public API, i.e., is any changed class annotated with
`@Public(Evolving)`: no
- The serializers: no
- The runtime per-record code paths (performance sensitive): yes
- Anything that affects deployment or recovery: JobManager (and its
components), Checkpointing, Kubernetes/Yarn, ZooKeeper: no
- The S3 file system connector: no
## Documentation
- Does this pull request introduce a new feature? no
- If yes, how is the feature documented? not applicable
--
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]
