jackjlli commented on PR #12297: URL: https://github.com/apache/pinot/pull/12297#issuecomment-1905150851
> Are we going to start an ExecutorService every time? The `releaseAndRemoveAllSegments` method is only invoked when shutting down the whole pinot-server. So within the lifecycle of a java process the executor will be started only once. > Did we take a thread dump to understand why the segment release was blocked / what was it waiting on? No we didn't take a thread dump but agree that we should have taken one. Basically, it timed out while waiting for the consumer thread to exit. The consumer thread may still be in `consumeLoop()` method, or the consumer thread was still busy processing the fetched events. But right now since the `stop()` method is called for each segment at a time, and the timeout for waiting for a segment to stop is 10 mins, it will make the whole server to wait for just 1 segment to stop, while other consuming segments are still functioning. Same for other tables in the same pinot-server that hasn't gotten the chance to stop. That's why this PR tries to invoke the stop call async so that all the segments of a table can be stopped together for server shutdown. -- 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: commits-unsubscr...@pinot.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
