Thesharing opened a new pull request #15500:
URL: https://github.com/apache/flink/pull/15500
## What is the purpose of the change
*This change fixes the incorrect way to calculate cross-region
ConsumedPartitionGroups in PipelinedRegionSchedulingStrategy in FLINK-21330. It
slows down the procedure of onExecutionStateChange, makes the complexity
increase from O(N) to O(N^2). Also the semantic of cross-region is totally
wrong.*
*To correctly calculate the cross-region ConsumedPartitionGroups, we can
just calculate the producer regions for all ConsumedPartitionGroups, and then
iterate all the regions and its ConsumedPartitionGroups. If the
ConsumedPartitionGroup has two or more producer regions, and the regions
contains the current region, it is a cross-region ConsumedPartitionGroup.*
*For more details, please check FLINK-22077*
## Brief change log
- *Fix the incorrect way to create cross-region ConsumedPartitionGroups in
PipelinedRegionSchedulingStrategy*
## Verifying this change
*This change is verified by existing tests in
PipelinedRegionSchedulingStrategyTest.
`testSchedulingTopologyWithCrossRegionConsumedPartitionGroups` tests the case
with cross-region ConsumedPartitionGroups illustrated in FLINK-22017. Also we
test this change in end-to-end tests.*
## Does this pull request potentially affect one of the following parts:
- Dependencies (does it add or upgrade a dependency): (yes / **no**)
- The public API, i.e., is any changed class annotated with
`@Public(Evolving)`: (yes / **no**)
- The serializers: (yes / **no** / don't know)
- The runtime per-record code paths (performance sensitive): (yes / **no**
/ don't know)
- Anything that affects deployment or recovery: JobManager (and its
components), Checkpointing, Kubernetes/Yarn/Mesos, ZooKeeper: (**yes** / no /
don't know)
- The S3 file system connector: (yes / **no** / don't know)
## Documentation
- Does this pull request introduce a new feature? (yes / **no**)
- If yes, how is the feature documented? (**not applicable** / docs /
JavaDocs / not documented)
--
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]
