SaurabhChawla100 commented on pull request #29413: URL: https://github.com/apache/spark/pull/29413#issuecomment-672981082
> > Using this VariableLinkedBlockingQueue, we are increasing certain percentage threshold for size of the queue at run time, Which is less aggressive compare to keeping the Integer.MAX_VALUE, Also I am planning to add the validation of spark driver memory to increase the size of the VariableLinkedBlockingQueue > > IIUC, `LinkedBlockingQueue` doesn't really allocate all the memory in once by setting the capacity to `Integer.MAX_VALUE`. > It also increase memory usage gradually at runtime. That's right, memory is allocated once the event is present in the Queue. In case when huge number of events flowing ,on keeping Integer.MAX_VALUE and there is no control on the number of events on the event queue, it keeps the large number of events queue(might be 100k or more), So there is chance of OOM is here if the size is not controlled and driver memory is not sufficient. VariableLinkedBlockingQueue is similar to LinkedBlockingQueue with the functionality of setting the size of the queue at run time. So we can make use of this VariableLinkedBlockingQueue and adjust the size based on various conditions ---------------------------------------------------------------- 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: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
