tgravescs commented on a change in pull request #27636:
[SPARK-30873][CORE][YARN]Handling Node Decommissioning for Yarn cluster manger
in Spark
URL: https://github.com/apache/spark/pull/27636#discussion_r390350870
##########
File path: core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala
##########
@@ -1525,10 +1539,25 @@ private[spark] class DAGScheduler(
s" ${task.stageAttemptId} and there is a more recent attempt for
that stage " +
s"(attempt ${failedStage.latestInfo.attemptNumber}) running")
} else {
+ if
(!event.reason.asInstanceOf[FetchFailed].countTowardsStageFailures) {
Review comment:
oh sorry, I misread it as countTowardsTaskFailures which is pre-existing.
Right now the logic for this is very specific to decommision, so think it
should be renamed to indicate so. If there are other cases this variable could
be useful - then the question is can you differenitate those from the
decomission case so you can still do this? If not just rename it to be specific
to decomission
----------------------------------------------------------------
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]
