otterc commented on code in PR #37533:
URL: https://github.com/apache/spark/pull/37533#discussion_r952800397
##########
core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala:
##########
@@ -2259,37 +2266,58 @@ private[spark] class DAGScheduler(
}
override def onShuffleMergeFailure(e: Throwable): Unit = {
+ if (e.isInstanceOf[IOException]) {
+ logInfo(s"Failed to connect external shuffle service
on " +
+ s"${shuffleServiceLoc.host} and add it to blacklist")
+
blockManagerMaster.removeShufflePushMergerLocation(shuffleServiceLoc.host)
+ finalizeBlackNodes.put(shuffleServiceLoc.host,
shuffleServiceLoc.host)
+ }
}
})
+
+ case _ =>
}
}
}, 0, TimeUnit.SECONDS)
} else {
- stage.shuffleDep.getMergerLocs.zipWithIndex.foreach {
- case (shuffleServiceLoc, index) =>
- // Sends async request to shuffle service to finalize shuffle
merge on that host
- // TODO: SPARK-35536: Cancel finalizeShuffleMerge if the stage is
cancelled
- // TODO: during shuffleMergeFinalizeWaitSec
- shuffleClient.finalizeShuffleMerge(shuffleServiceLoc.host,
- shuffleServiceLoc.port, shuffleId, shuffleMergeId,
- new MergeFinalizerListener {
- override def onShuffleMergeSuccess(statuses: MergeStatuses):
Unit = {
- assert(shuffleId == statuses.shuffleId)
- eventProcessLoop.post(RegisterMergeStatuses(stage,
MergeStatus.
- convertMergeStatusesToMergeStatusArr(statuses,
shuffleServiceLoc)))
- results(index).set(true)
- }
+ shuffleMergeFinalizeScheduler.schedule(new Runnable {
+ override def run(): Unit = {
+ stage.shuffleDep.getMergerLocs.zipWithIndex.foreach {
+ case (shuffleServiceLoc, index)
+ if finalizeBlackNodes.getIfPresent(shuffleServiceLoc.host) ==
null =>
+ // Sends async request to shuffle service to finalize shuffle
merge on that host
+ // TODO: SPARK-35536: Cancel finalizeShuffleMerge if the stage
is cancelled
+ // TODO: during shuffleMergeFinalizeWaitSec
+ shuffleClient.finalizeShuffleMerge(shuffleServiceLoc.host,
+ shuffleServiceLoc.port, shuffleId, shuffleMergeId,
+ new MergeFinalizerListener {
+ override def onShuffleMergeSuccess(statuses:
MergeStatuses): Unit = {
+ assert(shuffleId == statuses.shuffleId)
+ eventProcessLoop.post(RegisterMergeStatuses(stage,
MergeStatus.
+ convertMergeStatusesToMergeStatusArr(statuses,
shuffleServiceLoc)))
+ results(index).set(true)
+ }
- override def onShuffleMergeFailure(e: Throwable): Unit = {
- logWarning(s"Exception encountered when trying to finalize
shuffle " +
- s"merge on ${shuffleServiceLoc.host} for shuffle
$shuffleId", e)
- // Do not fail the future as this would cause dag scheduler
to prematurely
- // give up on waiting for merge results from the remaining
shuffle services
- // if one fails
- results(index).set(false)
- }
- })
- }
+ override def onShuffleMergeFailure(e: Throwable): Unit = {
+ logWarning(s"Exception encountered when trying to
finalize shuffle " +
+ s"merge on ${shuffleServiceLoc.host} for shuffle
$shuffleId", e)
+ // Do not fail the future as this would cause dag
scheduler to prematurely
+ // give up on waiting for merge results from the
remaining shuffle services
+ // if one fails
+ if (e.isInstanceOf[IOException]) {
+ logInfo(s"Failed to connect external shuffle service
on " +
+ s"${shuffleServiceLoc.host} and add it to blacklist")
+
blockManagerMaster.removeShufflePushMergerLocation(shuffleServiceLoc.host)
+ finalizeBlackNodes.put(shuffleServiceLoc.host,
shuffleServiceLoc.host)
+ }
+ results(index).set(false)
+ }
+ })
+
+ case (_, index) => results(index).set(true)
+ }
+ }
+ }, 0, TimeUnit.SECONDS)
Review Comment:
What happens when `shuffleMergeFinalizeScheduler` has just 1 thread?
--
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]
