Uniffle is a high performance, general purpose Remote Shuffle Service.

mapreduce remote-shuffle-service rss shuffle spark tez
8 Open Issues Need Help Last updated: Sep 11, 2025

Open Issues Need Help

View All on GitHub

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: The task is to implement a feature in Apache Uniffle to add a longer retry mechanism when a writer attempts to connect to the shuffle manager to obtain a shuffle handler. This is to improve the robustness of the system, especially under high GC pressure, where connection timeouts might occur.

Complexity: 4/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: The issue reports a NullPointerException (NPE) in the StatisticsCodec within the Apache Uniffle project. The reporter is willing to submit a pull request to fix the bug and has provided a screenshot illustrating the error. The task involves debugging the Uniffle codebase to identify the root cause of the NPE, implementing a fix, and submitting a pull request with thorough testing.

Complexity: 4/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: Implement a feature to display the compression ratio in the Uniffle + Spark UI tab. This involves adding code to calculate and display the compression ratio of shuffle data processed by Uniffle within the existing Spark UI.

Complexity: 4/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: The issue describes a NullPointerException (NPE) in the Apache Uniffle Spark client related to the shuffle push cost tracker. Debugging requires analyzing the provided screenshot and logs (currently empty) to identify the root cause of the NPE within the Uniffle Spark client code. This likely involves examining the code's handling of shuffle data pushing and its interaction with the cost tracker.

Complexity: 4/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: Debug a NullPointerException in the Apache Uniffle Spark UI rendering. The issue manifests as a broken UI tab, and requires investigation of Uniffle server and engine logs, configurations, and potentially code review to identify the root cause and provide a fix.

Complexity: 4/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: The task involves preventing chunk ID overflow in Apache Uniffle's ChunkCreator. This requires identifying the overflow condition in the provided code (link given in the issue description) and implementing a reset mechanism that sets the chunk ID to the next available ID in the chunk pool. A pull request is expected.

Complexity: 3/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez

AI Summary: The task is to improve the performance of the `reportShuffleResult` function in Apache Uniffle by replacing the current gRPC implementation with a Netty-based protocol. This aims to reduce CPU consumption during decoding.

Complexity: 4/5
good first issue

Uniffle is a high performance, general purpose Remote Shuffle Service.

Java
#mapreduce#remote-shuffle-service#rss#shuffle#spark#tez