walterddr commented on PR #10858:
URL: https://github.com/apache/pinot/pull/10858#issuecomment-1579274631
> so basically, there are dependencies between tasks from different stages,
and a fixed thread pool might have led to dead lock among them, like those
taking threads and running might actually wait for tasks waiting in the queue
to complete.
partially correct
- the worker/runner thread are deadlock/starvation free, so using a fixed
threadpool would not have any issues
- i modified them in this PR but it is not necessary. in fact in incline
to revert them b/c it help catch some distributed deadlock in workers.
- the server thread which handles compiling the query is not deadlock-free,
and it is designed specifically to be blocked there.
- which is the one that fixes the flaky test in this PR (the one I
commented: https://github.com/apache/pinot/pull/10858#discussion_r1220061034)
--
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]
