rmetzger opened a new pull request #15417:
URL: https://github.com/apache/flink/pull/15417


   ## What is the purpose of the change
   
   This commit fixes a test instability in 
ReactiveModeITCase.testScaleDownOnTaskManagerLoss() where a CountDownLatch was 
counted down on an execution of the test source that got cancelled for a scale 
up event (despite its name, a scale up can happen in the test as well). This 
test instability only happens if there is enough time between the notification 
of new slots so that the source can start in between. On fast machines, the 
execution gets cancelled during scheduling, because the slots arrive together.
   
   With this commit, we are not using a latch, but counting the number of 
running instances explicitly.
   
   


-- 
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.

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to