wsry opened a new pull request #15199:
URL: https://github.com/apache/flink/pull/15199
## What is the purpose of the change
Currently, sort-merge blocking shuffle implementation uses some unmanaged
direct memory and the netty thread for shuffle data read which can lead to
direct memory OOM and stress the netty thread with file IO. This patch
introduces a separated buffer pool and a separated thread pool for data read.
In the following patch, the problem will be totally solved by the IO scheduling
mechanism leveraging the introduced buffer pool and thread pool.
## Brief change log
- Introduce a separated buffer pool and a separated thread pool for
sort-merge blocking shuffle.
## Verifying this change
This change added tests BatchReadBufferPoolTest.
## 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:
us...@infra.apache.org
