Skip to content

[SPARK-20079][Core][yarn] Re registration of AM hangs spark cluster in yarn-client mode. #17480

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
wants to merge 4 commits into from
Closed

[SPARK-20079][Core][yarn] Re registration of AM hangs spark cluster in yarn-client mode. #17480

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
d1fe96a
SPARK-20079: Re registration of AM hangs spark cluster in yarn-client…
witgo Mar 30, 2017
2cc4dce
review commits
witgo Apr 4, 2017
013474a
review commits
witgo Apr 7, 2017
d3e69cf
Delete "initializing = true" in ExecutorAllocationManager.reset
witgo Apr 23, 2017
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
1 change: 0 additions & 1 deletion core/src/main/scala/org/apache/spark/ExecutorAllocationManager.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -249,7 +249,6 @@ private[spark] class ExecutorAllocationManager(
* yarn-client mode when AM re-registers after a failure.
*/
def reset(): Unit = synchronized {
initializing = true
Copy link
Contributor Author

Choose a reason for hiding this comment

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

@jerryshao @vanzin
I think that deleting the initializing = true is a good idea.

numExecutorsTarget = initialNumExecutors
numExecutorsToAdd = 1

Expand Down
16 changes: 16 additions & 0 deletions core/src/test/scala/org/apache/spark/ExecutorAllocationManagerSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -854,6 +854,17 @@ class ExecutorAllocationManagerSuite
assert(maxNumExecutorsNeeded(manager) === 1)
}

test("SPARK-20079: Re registration of AM hangs spark cluster in yarn-client mode") {
sc = createSparkContext()
val manager = sc.executorAllocationManager.get
assert(initializing(manager))
val stageInfo = createStageInfo(0, 2)
sc.listenerBus.postToAll(SparkListenerStageSubmitted(stageInfo))
assert(!initializing(manager))
manager.reset()
assert(!initializing(manager))
}

test("reset the state of allocation manager") {
sc = createSparkContext()
val manager = sc.executorAllocationManager.get
Expand Down Expand Up @@ -988,6 +999,7 @@ private object ExecutorAllocationManagerSuite extends PrivateMethodTester {
| Helper methods for accessing private methods and fields |
* ------------------------------------------------------- */

private val _initializing = PrivateMethod[Boolean]('initializing)
private val _numExecutorsToAdd = PrivateMethod[Int]('numExecutorsToAdd)
private val _numExecutorsTarget = PrivateMethod[Int]('numExecutorsTarget)
private val _maxNumExecutorsNeeded = PrivateMethod[Int]('maxNumExecutorsNeeded)
Expand All @@ -1011,6 +1023,10 @@ private object ExecutorAllocationManagerSuite extends PrivateMethodTester {
private val _localityAwareTasks = PrivateMethod[Int]('localityAwareTasks)
private val _hostToLocalTaskCount = PrivateMethod[Map[String, Int]]('hostToLocalTaskCount)

private def initializing(manager: ExecutorAllocationManager): Boolean = {
manager invokePrivate _initializing()
}

private def numExecutorsToAdd(manager: ExecutorAllocationManager): Int = {
manager invokePrivate _numExecutorsToAdd()
}
Expand Down