KeeProMise commented on PR #7287:
URL: https://github.com/apache/hadoop/pull/7287#issuecomment-2597676292

   
![image](https://github.com/user-attachments/assets/fc58d83a-c90d-4728-9377-bdc0a3f033aa)
   
![image](https://github.com/user-attachments/assets/490e9bc7-2bbb-4ce7-af19-a2a2c0f1f109)
   Hi, @hfutatzhanghb, I've thought about it, and here's a better approach:
   Firstly, the router manages its own code.
   The existing RouterRpcFairnessPolicyController can be utilized.
   Most crucially, when receiving a response from nn, the async responder may 
need to retry. At this time, the async responder will attempt to re-add the 
request to calls. If calls is full, it cannot be added, causing the async 
responder to hang. Even though when retrying, the async responder can hand over 
the request to the async handler for retry, the request still cannot be retried 
when calls is full. Retried requests should have a higher priority than the 
latest incoming requests because they were sent to the router earlier.


-- 
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