tgravescs commented on a change in pull request #27223:
[SPARK-30511][SPARK-28403][CORE] Don't treat failed/killed speculative tasks as
pending in Spark scheduler
URL: https://github.com/apache/spark/pull/27223#discussion_r371288191
##########
File path: core/src/main/scala/org/apache/spark/ExecutorAllocationManager.scala
##########
@@ -263,9 +263,15 @@ private[spark] class ExecutorAllocationManager(
*/
private def maxNumExecutorsNeeded(): Int = {
val numRunningOrPendingTasks = listener.totalPendingTasks +
listener.totalRunningTasks
- math.ceil(numRunningOrPendingTasks * executorAllocationRatio /
- tasksPerExecutorForFullParallelism)
- .toInt
+ val maxNeeded = math.ceil(numRunningOrPendingTasks *
executorAllocationRatio /
+ tasksPerExecutorForFullParallelism).toInt
+ if (listener.pendingSpeculativeTasks > 0 &&
tasksPerExecutorForFullParallelism > 1) {
+ // If we have pending speculative tasks, allocate one more executor to
satisfy the
+ // locality requirements of speculative tasks
+ maxNeeded + 1
Review comment:
another option would be to do the logic for speculative separately instead
of lumping it with the normal tasks. This might be an in between the 2
options. You could for instance calculate how many you need for regular tasks,
then calculate separately how many executors for speculative, if speculative
doesn't add another executor on top of the normal tasks then add one. That
would help with us double counting.
I'm fine with any of these as we can follow up to make it smarter in the
future.
----------------------------------------------------------------
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]
