Skip to content

[Hot Fix #469] Fix flaky test in SparkListenerSuite#516

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
andrewor14 wants to merge 1 commit into from
+3 −2
Closed

[Hot Fix #469] Fix flaky test in SparkListenerSuite #516

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 3 additions & 2 deletions core/src/test/scala/org/apache/spark/scheduler/SparkListenerSuite.scala
View file
Open in desktop
Original file line numberDiff line numberDiff line change
Expand Up@@ -50,9 +50,9 @@ class SparkListenerSuite extends FunSuite with LocalSparkContext with ShouldMatc
(1 to 5).foreach { _ => bus.post(SparkListenerJobEnd(0, JobSucceeded)) }
assert(counter.count === 0)

// Starting listener bus should flush all buffered events (asynchronously, hence the sleep)
// Starting listener bus should flush all buffered events
bus.start()
Thread.sleep(1000)
assert(bus.waitUntilEmpty(WAIT_TIMEOUT_MILLIS))
assert(counter.count === 5)

// After listener bus has stopped, posting events should not increment counter
Expand DownExpand Up@@ -177,6 +177,7 @@ class SparkListenerSuite extends FunSuite with LocalSparkContext with ShouldMatc
listener.stageInfos.clear()

rdd3.count()
assert(sc.listenerBus.waitUntilEmpty(WAIT_TIMEOUT_MILLIS))

@zsxwingzsxwingApr 24, 2014

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just realize that waitUntilEmpty is not enough. In waitUntilEmpty, it only checks eventQueue.isEmpty. But when eventQueue.isEmpty is true, there is still a chance that listeners do not finish their jobs or the memory is not synchronized.

@rxinrxinApr 24, 2014

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

cc @andrewor14

@andrewor14andrewor14Apr 24, 2014

Copy link
Copy Markdown
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch. A solution is to have eventQueue peek instead of take, such that we remove the event from the queue only after all listeners have finished processing it.

Though many other places have long been relying on waitUntilEmpty before this PR. It makes me wonder how likely this race condition actually causes a test failure... (nevertheless it's possible)

@andrewor14andrewor14Apr 24, 2014

Copy link
Copy Markdown
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hm, looks like LinkedBlockingQueue's take() is special in that it waits for the next item to be ready. It seems there isn't an equivalent for peek... we may have to synchronize some other way.

@zsxwingzsxwingApr 24, 2014

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Since waitUntilEmpty is only for test, is it possible that moving such wait logic to the SparkListener instances used in tests? E.g.,

importjava.util.concurrent.{CountDownLatch, TimeUnit}
classSaveStageAndTaskInfoextendsSparkListener {
valstageInfos= mutable.Map[StageInfo, Seq[(TaskInfo, TaskMetrics)]]()
vartaskInfoMetrics= mutable.Buffer[(TaskInfo, TaskMetrics)]()
vallatch=newCountDownLatch(1)
overridedefonTaskEnd(task: SparkListenerTaskEnd) {
valinfo= task.taskInfo
valmetrics= task.taskMetrics
if (info !=null&& metrics !=null) {
taskInfoMetrics += ((info, metrics))
}
}
overridedefonStageCompleted(stage: SparkListenerStageCompleted) {
stageInfos(stage.stageInfo) = taskInfoMetrics
taskInfoMetrics = mutable.Buffer.empty
latch.countDown()
}
defwaitForCompleted(timeoutMillis: Long) {
latch.await(timeoutMillis, TimeUnit.MILLISECONDS)
}
}

@andrewor14andrewor14Apr 25, 2014

Copy link
Copy Markdown
ContributorAuthor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@zsxwing we have to keep waitUntilEmpty in LiveListenerBus.scala (unfortunately) because DAGSchedulerSuite also uses it. I started a more general solution at #544. Maybe we can move the discussion there.

listener.stageInfos.size should be {2} // Shuffle map stage + result stage
val stageInfo3 = listener.stageInfos.keys.find(_.stageId == 2).get
stageInfo3.rddInfos.size should be {2} // ShuffledRDD, MapPartitionsRDD
Expand Down