maheshk114 commented on code in PR #52793:
URL: https://github.com/apache/spark/pull/52793#discussion_r2483170974
##########
core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala:
##########
@@ -975,7 +975,10 @@ class CoarseGrainedSchedulerBackend(scheduler:
TaskSchedulerImpl, val rpcEnv: Rp
val executorsToKill = knownExecutors
.filter { id => !executorsPendingToRemove.contains(id) }
.filter { id => force || !scheduler.isExecutorBusy(id) }
- executorsToKill.foreach { id => executorsPendingToRemove(id) =
!countFailures }
+ executorsToKill.foreach { id => {
Review Comment:
what about decommissioned executors ?
##########
core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala:
##########
@@ -589,7 +595,7 @@ private[spark] class TaskSchedulerImpl(
}
if (!launchedAnyTask) {
- taskSet.getCompletelyExcludedTaskIfAny(hostToExecutors).foreach {
taskIndex =>
+
taskSet.getCompletelyExcludedTaskIfAny(availableHostToExecutors).foreach {
taskIndex =>
Review Comment:
availableHostToExecutors or hostToAvailableExecutors
##########
core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala:
##########
@@ -589,7 +595,7 @@ private[spark] class TaskSchedulerImpl(
}
if (!launchedAnyTask) {
- taskSet.getCompletelyExcludedTaskIfAny(hostToExecutors).foreach {
taskIndex =>
+
taskSet.getCompletelyExcludedTaskIfAny(availableHostToExecutors).foreach {
taskIndex =>
Review Comment:
I think it will be simpler to just have a hash set updated for killed
executor and filter out from executorIdToHost
--
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]
