Daniel Darabos created SPARK-5102:
-------------------------------------
Summary: CompressedMapStatus needs to be registered with Kryo
Key: SPARK-5102
URL: https://issues.apache.org/jira/browse/SPARK-5102
Project: Spark
Issue Type: Bug
Affects Versions: 1.2.0
Reporter: Daniel Darabos
Priority: Minor
After upgrading from Spark 1.1.0 to 1.2.0 I got this exception:
{code}
Caused by: org.apache.spark.SparkException: Job aborted due to stage failure:
Task 0 in stage 0.0 failed 1 times, most recent failure: Lost task 0.0 in stage
0.0 (TID 0, localhost): java.lang.IllegalArgumentException: Class is not
registered: org.apache.spark.scheduler.CompressedMapStatus
Note: To register this class use:
kryo.register(org.apache.spark.scheduler.CompressedMapStatus.class);
at com.esotericsoftware.kryo.Kryo.getRegistration(Kryo.java:442)
at
com.esotericsoftware.kryo.util.DefaultClassResolver.writeClass(DefaultClassResolver.java:79)
at com.esotericsoftware.kryo.Kryo.writeClass(Kryo.java:472)
at com.esotericsoftware.kryo.Kryo.writeClassAndObject(Kryo.java:565)
at
org.apache.spark.serializer.KryoSerializerInstance.serialize(KryoSerializer.scala:165)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:206)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
at java.lang.Thread.run(Thread.java:745)
{code}
I had to register {{org.apache.spark.scheduler.CompressedMapStatus}} with Kryo.
I think this should be done in {{spark/serializer/KryoSerializer.scala}},
unless instances of this class are not expected to be sent over the wire.
(Maybe I'm doing something wrong?)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
