Github user srdo commented on the issue:
https://github.com/apache/storm/pull/2745
@revans2 Good point. It probably doesn't hurt much in the real code,
because the queue is only shut down when the worker or executor is shut down.
It could be an issue for tests with LocalCluster though.
Took a different approach. As far as I can tell, there isn't a great reason
that we need to interrupt the async loops through the JCQueue. We can just
interrupt the threads directly, which is already happening during
worker/executor shutdown.
I think the reason to interrupt via the queue is because the wait
strategies now include use of LockSupport.parkNanos, so there's no guarantee
that the executors will hit a Thread.sleep and throw an InterruptedException.
We can fix this by checking manually for interrupts, either in the async loop,
or in the wait strategies where we're calling parkNanos.
---
