advancedxy commented on PR #811:
URL: 
https://github.com/apache/incubator-uniffle/pull/811#issuecomment-1514882785

   > After rethinking, I think the spill should not block when acquiring the 
lock, because the addRecord can take a long time, especially for tight memory.
   > 
   > So in the latest commit. I use the `ReentrantLock#tryLock` to avoid this.
   
   Do you have any cases report that improves for the new fix? Seems that it 
would be quite possible that executor with memory pressure and high write 
frequency, the spill could not take effect due to that it cannot acquire the 
lock, which would  still causes executor OOM. 
   
   Do you think is it possible to adaptive increate try lock wait time to 
alleviate that cases?


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to