liupc commented on issue #24167: [SPARK-27214]Upgrading locality level when task set is starving URL: https://github.com/apache/spark/pull/24167#issuecomment-483088546 Yes, It's better to be fixed if we can find out a graceful way. Actually, as I mentioned in the PR description, if enabled dynamic allocaiton, the idle of the executors may cause executors to be removed and not reallocate new executors until the `numExecutorsTarget` is reset. This is really bad and even can not meet the point of `Letting other parallel TSM or job use resources with locality` Ref: https://github.com/apache/spark/blob/0bb716bac38488bc216fbda29ce54e93751e641b/core/src/main/scala/org/apache/spark/ExecutorAllocationManager.scala#L388 ``` // Do not request more executors if it would put our target over the upper bound if (numExecutorsTarget >= maxNumExecutors) { logDebug(s"Not adding executors because our current target total " + s"is already $numExecutorsTarget (limit $maxNumExecutors)") numExecutorsToAdd = 1 return 0 } ```
---------------------------------------------------------------- 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]
