wsry opened a new pull request #18470:
URL: https://github.com/apache/flink/pull/18470
## What is the purpose of the change
Currently, for sort-shuffle, the maximum number of buffers can be used per
result partition is Integer.MAX_VALUE. However, if too many buffers are taken
by one result partition, other result partitions and input gates may spend too
much time waiting for buffers which can influence performance. This ticket aims
to restrict the maximum number of buffers can be used per result partition and
the selected value is an empirical one based on the TPC-DS test results.
## Brief change log
- Use Math.max(min, 4 * numSubpartitions) as the maximum number of buffers
can be used per result partition for sort-shuffle.
## Verifying this change
This change is already covered by existing 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, 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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
