pan3793 commented on PR #38902: URL: https://github.com/apache/spark/pull/38902#issuecomment-1341265075
@tedyu sorry if my words is ambiguous, the "listener" means `org.apache.spark.scheduler.SparkLisenter`, not the subscribers of "kubernetes executor snapshots". During shutdown process, they share the "hadoop.service.shutdown.timeout", increasing KUBERNETES_EXECUTOR_SNAPSHOTS_SUBSCRIBERS_GRACE_PERIOD means decreasing grace period of `listenerBus.stop()`. https://github.com/apache/spark/blob/1d2159f888139e65c80db2003d521b0f684df83a/core/src/main/scala/org/apache/spark/SparkContext.scala#L2125-L2136 For our cases, the authors of listener may not know much about the Spark internal, when they increases 20s to `hadoop.service.shutdown.timeout`, they may expect that the listener will get another 20s grace period. To be clear, I am not oppose to your idea, just left my concerns. -- 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]
