HeartSaVioR edited a comment on issue #25853: [SPARK-21869][SS] Apply Apache Commons Pool to Kafka producer URL: https://github.com/apache/spark/pull/25853#issuecomment-548644351 That was just "one node" setup. That's intentional because network latency could affect the result, especially we focus to test Kafka producer which would not be CPU-intensive. * Spark ran with standalone mode, default configuration. * ZK and Kafka ran from same node with Spark. Kafka 2.3.1, ZK is just what Kafka contained. * Note that I don't replace kafka-client artifact. Just pulled all transitive dependencies via `--packages`. * 1 broker with foreground run. default configuration. ZK was also run with foreground. Note that my goal of experiment is not proving the patch outperforms than current. My goal is showing that the patch doesn't hurt performance, or at least the impact is trivial. If we really want to have strict performance test to compare both we may want to have dedicated machine to run them with isolation.
---------------------------------------------------------------- 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] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
