summaryzb opened a new issue, #1264:
URL: https://github.com/apache/incubator-uniffle/issues/1264

   ### Code of Conduct
   
   - [X] I agree to follow this project's [Code of 
Conduct](https://www.apache.org/foundation/policies/conduct)
   
   
   ### Search before asking
   
   - [X] I have searched in the 
[issues](https://github.com/apache/incubator-uniffle/issues?q=is%3Aissue) and 
found no similar issues.
   
   
   ### Describe the feature
   
   When task is killed for stage cancel, another task attempt succeed or some 
other reasons, The `AddBlockEvent` handling and `sendShuffleData`  still work. 
   Although `needCancelRequest` may cancel some work, but the `AddBlockEvent` 
in the blocking queue of threadPool still holds the shuffleblockdata, and so as 
to the rpc request that are already called but waiting for repsonse.
   
   That will cause 3 problems:
   1. We freeAll memory onece the task is killed, but the shuffleBlockData hold 
by the async thread still occupy memory
   2. Many useless runnable related to the kille task are still working or wait 
to be executed
   3. Currently`checkBlockSendResult` can not be interrupted
   
   
   ### Motivation
   
   _No response_
   
   ### Describe the solution
   
   1. Cancel  all the runnable that are wait to be executed or  blocking in 
waiting for rpc callback
   2. Interrupt `checkBlockSendResult` immediately
   
   ### Additional context
   
   _No response_
   
   ### Are you willing to submit PR?
   
   - [X] Yes I am willing to submit a PR!


-- 
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