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-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
