PatrickRen opened a new pull request #15461:
URL: https://github.com/apache/flink/pull/15461
## What is the purpose of the change
This pull request removes reader id to split assignment mapping from state,
and change the logic of worker coordinator thread to avoid double-bookkeeping
discovered partitions in ```KafkaSourceEnumerator```.
## Brief change log
*(for example:)*
- Change ```KafkaSourceEnumState``` and its serializer to store assigned
partitions instead of split assignments, since split assignments might change
after job failover.
- Change the logic of worker thread and coordinator thread. Worker thread
fetches all topic information from Kafka cluster, and coordinator thread checks
if there's any partition changes. This solves potential inconsistency between
two sets of discovered partitions kept separately by worker and coordinator
thread.
## Verifying this change
This change added tests and can be verified as follows:
- Added test for the new serializer of ```KafkaSourceEnumState```.
Backward compatibility is also tested.
- Logic of worker and coordinator thread has been tested by existing tests.
## 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: (yes)
- The runtime per-record code paths (performance sensitive): (no)
- Anything that affects deployment or recovery: JobManager (and its
components), Checkpointing, Kubernetes/Yarn/Mesos, 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.
For queries about this service, please contact Infrastructure at:
[email protected]
