Thesharing opened a new pull request #16538:
URL: https://github.com/apache/flink/pull/16538
## What is the purpose of the change
*This pull request is the second part of optimizing the deployment of tasks.
In this pull request, we propose to distribute the cached ShuffleDescriptors
via the blob server if their overall size has exceeded the threshold. For more
details please check FLINK-23218. The first part of the optimization is located
at #16314.*
*In this pull request we remove the cache from blob server on JobManager
once it's no longer required. To remove the cache from the blob cache located
on TaskManager, please check #16498.*
## Brief change log
- *The cached ShuffleDescriptors are stored in the blob server if its size
exceeded the threshold.*
- *The cached ShuffleDescriptors will be removed from the blob server once
it's no longer used on JobManager.*
- *TaskManagers can retrieve the ShuffleDescriptors from the blob cache.*
## Verifying this change
- *Added unit tests to test the ShuffleDescriptors are cached and removed
correctly.*
- *Added unit tests to verify that deleting files on blob server works as
expected.*
- *Manually verified the change by running an end-to-end test. The job has
two JobVertices. One is source, while the other is sink. The parallelism is 8k
and 16k. They are connected with an all-to-all edge. Both streaming mode and
batch mode are tested.*
## 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]
