AngersZhuuuu commented on code in PR #52793:
URL: https://github.com/apache/spark/pull/52793#discussion_r2567008129
##########
core/src/main/scala/org/apache/spark/scheduler/TaskSchedulerImpl.scala:
##########
@@ -605,7 +611,8 @@ private[spark] class TaskSchedulerImpl(
// If there are no idle executors and dynamic allocation is
enabled, then we would
// notify ExecutorAllocationManager to allocate more executors
to schedule the
// unschedulable tasks else we will abort immediately.
- executorIdToRunningTaskIds.find(x => !isExecutorBusy(x._1))
match {
+ executorIdToRunningTaskIds.find(
+ x => isExecutorAvailable(x._1) && !isExecutorBusy(x._1)) match
{
Review Comment:
Since here we can't use `hostToExecutors`, if use `hostToExecutors`, will
won't find out excluded tasks since `pendingToRemoveExecutors` not excluded. So
we can't use `hostToExecutors` when call
`taskSet.getCompletelyExcludedTaskIfAny()`. if we extend `isExecutorBusy()`, we
will add more check rule in `isExecutorBusy`, it's also called in
`CoarseGrainedSchedulerBackend`.`killExecutors()`, we add more check, so I
think not necessary.....
--
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]
