apache / incubator-uniffle

Uniffle is a high performance, general purpose Remote Shuffle Service.
https://uniffle.apache.org/
Apache License 2.0
382 stars 149 forks source link

[Bug] Occasionally client sent blocks to server time out. #834

Open bin41215 opened 1 year ago

bin41215 commented 1 year ago

Code of Conduct

Search before asking

Describe the bug

tez writes data to the server (refer to mr's method), and occasionally sends data time out. (10min)

image

Affects Version(s)

master

Uniffle Server Log Output

No response

Uniffle Engine Log Output

No response

Uniffle Server Configurations

No response

Uniffle Engine Configurations

No response

Additional context

No response

Are you willing to submit PR?

zuston commented 1 year ago

The detailed stacktrace is missing, could you help attach more?

BTW, one possible problem is the server GC. But this should be confirmed by more info.

bin41215 commented 1 year ago

The detailed stacktrace is missing, could you help attach more?

BTW, one possible problem is the server GC. But this should be confirmed by more info.

Sorry, I was delayed by other things these days. This problem seems to be the problem of shuffleId when the tez client is implemented. It has been modified here, and it will be observed for a few days.

lifeSo commented 10 months ago

You can update to the latest uniffle version. And is this problem show again ?