Skip to content

Conversation

@tedyu
Copy link
Contributor

@tedyu tedyu commented Nov 12, 2015

Here is the stack trace:

java.lang.NullPointerException
      at org.apache.spark.shuffle.sort.SortShuffleManager.registerShuffle(SortShuffleManager.scala:91)
      at org.apache.spark.ShuffleDependency.<init>(Dependency.scala:90)
      at org.apache.spark.rdd.ShuffledRDD.getDependencies(ShuffledRDD.scala:80)
      at org.apache.spark.rdd.RDD$$anonfun$dependencies$2.apply(RDD.scala:226)
      at org.apache.spark.rdd.RDD$$anonfun$dependencies$2.apply(RDD.scala:224)
      at scala.Option.getOrElse(Option.scala:120)
      at org.apache.spark.rdd.RDD.dependencies(RDD.scala:224)
      at org.apache.spark.scheduler.DAGScheduler.visit$1(DAGScheduler.scala:386)
      at org.apache.spark.scheduler.DAGScheduler.getParentStages(DAGScheduler.scala:398)
      at org.apache.spark.scheduler.DAGScheduler.getParentStagesAndId(DAGScheduler.scala:299)
      at org.apache.spark.scheduler.DAGScheduler.newResultStage(DAGScheduler.scala:334)
      at org.apache.spark.scheduler.DAGScheduler.handleJobSubmitted(DAGScheduler.scala:836)
      at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.doOnReceive(DAGScheduler.scala:1607)
      at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1599)
      at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1588)

@andrewor14
Copy link
Contributor

@tdas @zsxwing

@zsxwing
Copy link
Member

zsxwing commented Nov 12, 2015

I don't think it's worth to fix it here. This failure is because TrackStateRDDSuite creates its SparkContext in the constructor. For maven build, it will create all suites at the beginning. So TrackStateRDDSuite's SparkEnv will be set to null when some test before TrackStateRDDSuite stops another SparkContext. See #9668

@tedyu
Copy link
Contributor Author

tedyu commented Nov 12, 2015

I think both fixes should be kept.

@zsxwing
Copy link
Member

zsxwing commented Nov 12, 2015

If SparkEnv is null, I think it will crash somewhere as well. So why we need to fix it here?

@andrewor14
Copy link
Contributor

I agree. It seems that this patch is a bandaid without addressing the real issue. Once we have #9668 we don't need this patch.

@tedyu tedyu closed this Nov 12, 2015
@SparkQA
Copy link

SparkQA commented Nov 12, 2015

Test build #45763 has finished for PR 9672 at commit 0025e2b.

  • This patch passes all tests.
  • This patch merges cleanly.
  • This patch adds no public classes.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants