[SPARK-17611] YarnShuffleServiceSuite swallows exceptions, doesn't really test a few things - ASF JIRA

Attach files

Attach Screenshot

Voters

Watch issue

Watchers

Create sub-task

Link

Clone

Update Comment Author

Replace String in Comment

Update Comment Visibility

Delete Comments

XML

Word

Printable

JSON

Details

Type: Bug
Status: Resolved
Priority: Minor
Resolution: Fixed
Affects Version/s: None
Fix Version/s: 2.1.0
Component/s: Spark Core, Tests, YARN
Labels:
None

Description

If you look at the logs, you'll see exceptions like this:

16/09/20 10:03:45.081 pool-1-thread-1-ScalaTest-running-YarnShuffleServiceSuite ERROR YarnShuffleService: Exception when initializing application application_0_0001
java.lang.NullPointerException
        at io.netty.buffer.Unpooled.wrappedBuffer(Unpooled.java:183)
        at org.apache.spark.network.util.JavaUtils.bytesToString(JavaUtils.java:80)
        at org.apache.spark.network.sasl.ShuffleSecretManager.registerApp(ShuffleSecretManager.java:62)
        at org.apache.spark.network.yarn.YarnShuffleService.initializeApplication(YarnShuffleService.java:254)
        at org.apache.spark.network.yarn.YarnShuffleServiceSuite$$anonfun$1.apply$mcV$sp(YarnShuffleServiceSuite.scala:89)
        at org.apache.spark.network.yarn.YarnShuffleServiceSuite$$anonfun$1.apply(YarnShuffleServiceSuite.scala:81)
        at org.apache.spark.network.yarn.YarnShuffleServiceSuite$$anonfun$1.apply(YarnShuffleServiceSuite.scala:81)

The underlying reason is that the suite is testing with auth on, but providing a null auth secret. But because of catch blocks in the shuffle service code, and failure to properly check things in the suite, the tests pass even though the underlying code is failing.