chirag-wadhwa5 commented on PR #19405:
URL: https://github.com/apache/kafka/pull/19405#issuecomment-2783292100

   Thanks for the review. Answering the questions -
   1. The flakiness could not be reproduced locally in the first place, so 
there is no proof of that. Honestly reproducing the error atleast once locally 
would be extremely helpful, but that hasn't been possible yet.
   2. Maybe I could change it to 1 second. The idea was that since there is 
some bundling up of requests in the timingWheel, and the clock is advanced 
every 200ms, maybe the acquisition lock timer task would be expired in the 
second round of clock advance (after 400ms), but the waitFor condition only 
waits for 300ms, so it never gets the opportunity to get expired. So the 
timeout should be something greater than 400ms plus the time it takes to 
execute the expiration logic. But the time we are saving by reducing it from 2 
seconds to 500ms is just 5 to 10 seconds, which I don't think should be any 
cause of concern. So if it's not a big problem, let's keep it 2 seconds to be 
absolutely sure. Otherwise I can reduce it to 1 second.


-- 
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]

Reply via email to