Skip to content

[#2529] fix(spark3): Incorrect clientInfo without nettyPort if netty is enabled #2530

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jun 27, 2025

Conversation

zuston
Copy link
Member

@zuston zuston commented Jun 27, 2025

What changes were proposed in this pull request?

If the shuffle server enables Netty RPC, clients collect push cost metrics along with client information. However, the collected shuffle server information lacks the Netty port, which causes subsequent analysis to be inaccurate or invalid.

Why are the changes needed?

for #2529
image

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Yes. Test in the internal cluster spark jobs and existing UTs

@jerqi
Copy link
Contributor

jerqi commented Jun 27, 2025

Could you add a UT for this if this is a fix?

Copy link

Test Results

 3 049 files  ±0   3 049 suites  ±0   6h 48m 47s ⏱️ + 1m 10s
 1 188 tests ±0   1 187 ✅ ±0   1 💤 ±0  0 ❌ ±0 
15 067 runs  ±0  15 052 ✅ ±0  15 💤 ±0  0 ❌ ±0 

Results for commit b6622fe. ± Comparison against base commit 887d042.

@zuston
Copy link
Member Author

zuston commented Jun 27, 2025

Could you add a UT for this if this is a fix?

I really want to add UTs but it's hard to simulate

@jerqi
Copy link
Contributor

jerqi commented Jun 27, 2025

You just need to add a UT to test constructor of ShuffleServerGrpcClient.java.

@zuston
Copy link
Member Author

zuston commented Jun 27, 2025

You just need to add a UT to test constructor of ShuffleServerGrpcClient.java.

Aha. It's easy but not effective. And this have been tested in current codebase

@zuston
Copy link
Member Author

zuston commented Jun 27, 2025

Thanks @jerqi . Merged.

@zuston zuston merged commit f194cc5 into apache:master Jun 27, 2025
41 checks passed
@zuston zuston deleted the clientInfoBug branch June 27, 2025 11:08
zuston added a commit to zuston/incubator-uniffle that referenced this pull request Jul 2, 2025
…netty is enabled (apache#2530)

### What changes were proposed in this pull request?

If the shuffle server enables Netty RPC, clients collect push cost metrics along with client information. However, the collected shuffle server information lacks the Netty port, which causes subsequent analysis to be inaccurate or invalid.

### Why are the changes needed?

for apache#2529 
![image](https://github.com/user-attachments/assets/0c550086-ffeb-43c8-8fc2-3a2e42e3150e)


### Does this PR introduce _any_ user-facing change?

No.

### How was this patch tested?

Yes. Test in the internal cluster spark jobs and existing UTs
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants