vanzin commented on a change in pull request #23951: [SPARK-13704][CORE][YARN]
Re-implement RackResolver to reduce resolving time
URL: https://github.com/apache/spark/pull/23951#discussion_r268348986
##########
File path: core/src/main/scala/org/apache/spark/scheduler/TaskSetManager.scala
##########
@@ -186,8 +186,23 @@ private[spark] class TaskSetManager(
// Add all our tasks to the pending lists. We do this in reverse order
// of task index so that tasks with low indices get launched first.
- for (i <- (0 until numTasks).reverse) {
- addPendingTask(i)
+ addPendingTasks()
+
+ private def addPendingTasks(): Unit = {
+ val (_, duration) = Utils.timeTakenMs {
+ val hostToIndices = new HashMap[String, ArrayBuffer[Int]]()
+ for (i <- (0 until numTasks).reverse) {
+ addPendingTask(i, initializing = true, Some(hostToIndices))
+ }
+ // Resolve the rack for each host. This can be slow, so de-dupe the list
of hosts,
+ // and assign the rack to all relevant task indices.
+ for (
+ (Some(rack), indices) <-
sched.getRacksForHosts(hostToIndices.keySet.toSeq)
+ .zip(hostToIndices.values)) {
+ pendingTasksForRack.getOrElseUpdate(rack, new ArrayBuffer) ++= indices
+ }
+ }
+ logDebug(s"Adding pending tasks take $duration ms")
Review comment:
s/take/took
----------------------------------------------------------------
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]
