[jira] [Created] (SPARK-40582) NullPointerException: Cannot invoke invalidateSerializedMapOutputStatusCache() because "shuffleStatus" is null

["Garret Wilson (Jira)" <ji...@apache.org>]

Garret Wilson created SPARK-40582:
-------------------------------------

             Summary: NullPointerException: Cannot invoke invalidateSerializedMapOutputStatusCache() because "shuffleStatus" is null
                 Key: SPARK-40582
                 URL: https://issues.apache.org/jira/browse/SPARK-40582
             Project: Spark
          Issue Type: Bug
          Components: Spark Core
    Affects Versions: 3.3.0
            Reporter: Garret Wilson


I'm running a simple little Spark 3.3.0 pipeline on Windows 10 using Java 17 and UDFs. I hardly do anything interesting, and now when I run the pipeline on only 30,000 records I'm getting this:

{noformat}
[ERROR] Error in removing shuffle 2
java.lang.NullPointerException: Cannot invoke "org.apache.spark.ShuffleStatus.invalidateSerializedMapOutputStatusCache()" because "shuffleStatus" is null
        at org.apache.spark.MapOutputTrackerMaster.$anonfun$unregisterShuffle$1(MapOutputTracker.scala:882)
        at org.apache.spark.MapOutputTrackerMaster.$anonfun$unregisterShuffle$1$adapted(MapOutputTracker.scala:881)
        at scala.Option.foreach(Option.scala:437)
        at org.apache.spark.MapOutputTrackerMaster.unregisterShuffle(MapOutputTracker.scala:881)
        at org.apache.spark.storage.BlockManagerStorageEndpoint$$anonfun$receiveAndReply$1.$anonfun$applyOrElse$3(BlockManagerStorageEndpoint.scala:59)
        at scala.runtime.java8.JFunction0$mcZ$sp.apply(JFunction0$mcZ$sp.scala:17)
        at org.apache.spark.storage.BlockManagerStorageEndpoint.$anonfun$doAsync$1(BlockManagerStorageEndpoint.scala:89)
        at scala.concurrent.Future$.$anonfun$apply$1(Future.scala:678)
        at scala.concurrent.impl.Promise$Transformation.run(Promise.scala:467)
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
        at java.base/java.lang.Thread.run(Thread.java:833)
{noformat}

I searched and couldn't find any of the principal terms in the error message.

Disconcerting that Spark is breaking at what seems to be a fundamental part of processing, and with a {{NullPointerException}} at that.

_Update:_ Now it just happened with only 1000 records. But then I reran the pipeline immediately with no changes, and it succeeded. So this {{NullPointerException}} bug is nondeterministic. Not good at all.

I have already asked this question on [Stack Overflow|https://stackoverflow.com/q/73732970], and even posted a bounty, with no solutions. (The only answer so far is from someone who doesn't even use Spark and just posted links.)



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org