bmarcott commented on a change in pull request #27207: [WIP][SPARK-18886][CORE] Make Locality wait time measure resource under utilization due to delay scheduling. URL: https://github.com/apache/spark/pull/27207#discussion_r373821397
########## File path: core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala ########## @@ -100,6 +101,11 @@ private[spark] class TaskSchedulerImpl( // on this class. Protected by `this` private val taskSetsByStageIdAndAttempt = new HashMap[Int, HashMap[Int, TaskSetManager]] + // keyed by task set stage id + // value is true if there have been no resources rejected due to delay scheduling + // since the last "full" resource offer + private val noDelayScheduleRejects = new mutable.HashMap[Int, Boolean]() Review comment: what do you see as pros/cons of keeping this map vs putting in TSM? The `TaskSchedulerImpl` is the only one setting/updating the variable based on knowledge only it has (no schedule delay reject). ---------------------------------------------------------------- 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]
