[SPARK-13909] DataFrames DISK_ONLY persistence leads to OOME

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Incomplete
• Affects Version/s: 1.6.0
• Component/s: Spark Core
• Labels: bulk-closed dataframe

Environment

debian:jessie, java 1.8, hadoop 2.6.0, current zeppelin snapshot

• Assignee: Unassigned
• Reporter: Jakub Liska
• Created: 15/Mar/16 13:04
• Updated: 21/May/19 04:33
• Resolved: 21/May/19 04:33

Description

Hey, I migrated to 1.6.0, and suddenly `persist` behaves as if it was `MEMORY_ONLY` instead of `DISK_ONLY` so that it eventually ends with OOME. However if I remove `persist` it works fine. I'm calling this snippet from Zeppelin notebook :

val coreRdd = sc.textFile("s3n://gwiq-views-p/external/core/tsv/*.tsv").map(_.split("\t")).map( fields => Row(fields:_*) )
val coreDataFrame = sqlContext.createDataFrame(coreRdd, schema)
coreDataFrame.registerTempTable("core")
coreDataFrame.persist(StorageLevel.DISK_ONLY)

SELECT COUNT(*) FROM core

------ Create new SparkContext spark://master:7077 -------
Exception in thread "pool-1-thread-5" java.lang.OutOfMemoryError: GC overhead limit exceeded
	at com.google.gson.internal.bind.ObjectTypeAdapter.read(ObjectTypeAdapter.java:66)
	at com.google.gson.internal.bind.ObjectTypeAdapter.read(ObjectTypeAdapter.java:69)
	at com.google.gson.internal.bind.TypeAdapterRuntimeTypeWrapper.read(TypeAdapterRuntimeTypeWrapper.java:40)
	at com.google.gson.internal.bind.MapTypeAdapterFactory$Adapter.read(MapTypeAdapterFactory.java:188)
	at com.google.gson.internal.bind.MapTypeAdapterFactory$Adapter.read(MapTypeAdapterFactory.java:146)
	at com.google.gson.Gson.fromJson(Gson.java:791)
	at com.google.gson.Gson.fromJson(Gson.java:757)
	at com.google.gson.Gson.fromJson(Gson.java:706)
	at org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer.convert(RemoteInterpreterServer.java:417)
	at org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer.getProgress(RemoteInterpreterServer.java:384)
	at org.apache.zeppelin.interpreter.thrift.RemoteInterpreterService$Processor$getProgress.getResult(RemoteInterpreterService.java:1376)
	at org.apache.zeppelin.interpreter.thrift.RemoteInterpreterService$Processor$getProgress.getResult(RemoteInterpreterService.java:1361)
	at org.apache.thrift.ProcessFunction.process(ProcessFunction.java:39)
	at org.apache.thrift.TBaseProcessor.process(TBaseProcessor.java:39)
	at org.apache.thrift.server.TThreadPoolServer$WorkerProcess.run(TThreadPoolServer.java:285)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
	at java.lang.Thread.run(Thread.java:745)

I'm using this https://github.com/gettyimages/docker-spark setup with a zeppelin docker container...

    ZEPPELIN_JAVA_OPTS: -Dspark.executor.memory=16g -Dspark.serializer=org.apache.spark.serializer.KryoSerializer -Dspark.app.id=zeppelin
    SPARK_SUBMIT_OPTIONS: --driver-memory 1g --repositories https://oss.sonatype.org/content/repositories/snapshots --packages com.viagraphs:spark-extensions_2.10:1.04-SNAPSHOT --jars=file:/usr/spark-1.6.0-bin-hadoop2.6/lib/aws-java-sdk-1.7.14.jar,file:/usr/spark-1.6.0-bin-hadoop2.6/lib/hadoop-aws-2.6.0.jar,file:/usr/spark-1.6.0-bin-hadoop2.6/lib/google-collections-1.0.jar,file:/usr/spark-1.6.0-bin-hadoop2.6/lib/joda-time-2.8.2.jar
    SPARK_WORKER_CORES: 8
    SPARK_WORKER_MEMORY: 16g