Spark Java Error: Size exceeds Integer.MAX_VALUE
I am trying to use spark for some simple machine learning task. I used pyspark and spark 1.2.0 to do a simple logistic regression problem. I have 1.2 million records for training, and I hashed the features of the records. When I set the number of hashed features as 1024, the program works fine, but when I set the number of hashed features as 16384, the program fails several times with the following error:
Py4JJavaError: An error occurred while calling o84.trainLogisticRegressionModelWithSGD.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 1 in stage 4.0 failed 4 times, most recent failure: Lost task 1.3 in stage 4.0 (TID 9, workernode0.sparkexperience4a7.d5.internal.cloudapp.net): java.lang.RuntimeException: java.lang.IllegalArgumentException: Size exceeds Integer.MAX_VALUE
at sun.nio.ch.FileChannelImpl.map(FileChannelImpl.java:828)
at org.apache.spark.storage.DiskStore.getBytes(DiskStore.scala:123)
at org.apache.spark.storage.DiskStore.getBytes(DiskStore.scala:132)
at org.apache.spark.storage.BlockManager.doGetLocal(BlockManager.scala:517)
at org.apache.spark.storage.BlockManager.getBlockData(BlockManager.scala:307)
at org.apache.spark.network.netty.NettyBlockRpcServer$$anonfun$2.apply(NettyBlockRpcServer.scala:57)
at org.apache.spark.network.netty.NettyBlockRpcServer$$anonfun$2.apply(NettyBlockRpcServer.scala:57)
at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244)
at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:244)
at scala.collection.IndexedSeqOptimized$class.foreach(IndexedSeqOptimized.scala:33)
at scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:108)
at scala.collection.TraversableLike$class.map(TraversableLike.scala:244)
at scala.collection.mutable.ArrayOps$ofRef.map(ArrayOps.scala:108)
at org.apache.spark.network.netty.NettyBlockRpcServer.receive(NettyBlockRpcServer.scala:57)
at org.apache.spark.network.server.TransportRequestHandler.processRpcRequest(TransportRequestHandler.java:124)
at org.apache.spark.network.server.TransportRequestHandler.handle(TransportRequestHandler.java:97)
at org.apache.spark.network.server.TransportChannelHandler.channelRead0(TransportChannelHandler.java:91)
at org.apache.spark.network.server.TransportChannelHandler.channelRead0(TransportChannelHandler.java:44)
at io.netty.channel.SimpleChannelInboundHandler.channelRead(SimpleChannelInboundHandler.java:105)
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333)
at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:319)
at io.netty.handler.codec.MessageToMessageDecoder.channelRead(MessageToMessageDecoder.java:103)
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333)
at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:319)
at io.netty.handler.codec.ByteToMessageDecoder.channelRead(ByteToMessageDecoder.java:163)
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333)
at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:319)
at io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:787)
at io.netty.channel.nio.AbstractNioByteChannel$NioByteUnsafe.read(AbstractNioByteChannel.java:130)
at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:511)
at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:468)
at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:382)
at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:354)
at io.netty.util.concurrent.SingleThreadEventExecutor$2.run(SingleThreadEventExecutor.java:116)
at java.lang.Thread.run(Thread.java:745)
at org.apache.spark.network.client.TransportResponseHandler.handle(TransportResponseHandler.java:156)
at org.apache.spark.network.server.TransportChannelHandler.channelRead0(TransportChannelHandler.java:93)
at org.apache.spark.network.server.TransportChannelHandler.channelRead0(TransportChannelHandler.java:44)
at io.netty.channel.SimpleChannelInboundHandler.channelRead(SimpleChannelInboundHandler.java:105)
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333)
at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:319)
at io.netty.handler.codec.MessageToMessageDecoder.channelRead(MessageToMessageDecoder.java:103)
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333)
at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:319)
at io.netty.handler.codec.ByteToMessageDecoder.channelRead(ByteToMessageDecoder.java:163)
at io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:333)
at io.netty.channel.AbstractChannelHandlerContext.fireChannelRead(AbstractChannelHandlerContext.java:319)
at io.netty.channel.DefaultChannelPipeline.fireChannelRead(DefaultChannelPipeline.java:787)
at io.netty.channel.nio.AbstractNioByteChannel$NioByteUnsafe.read(AbstractNioByteChannel.java:130)
at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:511)
at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:468)
at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:382)
at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:354)
at io.netty.util.concurrent.SingleThreadEventExecutor$2.run(SingleThreadEventExecutor.java:116)
at java.lang.Thread.run(Thread.java:745)
Driver stacktrace:
at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$failJobAndIndependentStages(DAGScheduler.scala:1214)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1203)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1202)
at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59)
at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:47)
at org.apache.spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:1202)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:696)
at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:696)
at scala.Option.foreach(Option.scala:236)
at org.apache.spark.scheduler.DAGScheduler.handleTaskSetFailed(DAGScheduler.scala:696)
at org.apache.spark.scheduler.DAGSchedulerEventProcessActor$$anonfun$receive$2.applyOrElse(DAGScheduler.scala:1420)
at akka.actor.Actor$class.aroundReceive(Actor.scala:465)
at org.apache.spark.scheduler.DAGSchedulerEventProcessActor.aroundReceive(DAGScheduler.scala:1375)
at akka.actor.ActorCell.receiveMessage(ActorCell.scala:516)
at akka.actor.ActorCell.invoke(ActorCell.scala:487)
at akka.dispatch.Mailbox.processMailbox(Mailbox.scala:238)
at akka.dispatch.Mailbox.run(Mailbox.scala:220)
at akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinTask.exec(AbstractDispatcher.scala:393)
at scala.concurrent.forkjoin.ForkJoinTask.doExec(ForkJoinTask.java:260)
at scala.concurrent.forkjoin.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:1339)
at scala.concurrent.forkjoin.ForkJoinPool.runWorker(ForkJoinPool.java:1979)
at scala.concurrent.forkjoin.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:107)
This error happens when I training the LogisticRegressionWithSGD after transferring the data into LabeledPoint.
Does anyone have a idea on this?
My code is as follows (I am using a IPython Notebook for this):
from pyspark.mllib.regression import LabeledPoint
from pyspark.mllib.classification import LogisticRegressionWithSGD
from numpy import array
from sklearn.feature_extraction import FeatureHasher
from pyspark import SparkContext
sf = SparkConf().setAppName("test").set("spark.executor.memory", "50g").set("spark.cores.max", 30)
sc = SparkContext(conf=sf)
training_file = sc.textFile("train_small.txt")
def hash_feature(line):
values = [0, dict()]
for index, x in enumerate(line.strip("\n").split('\t')):
if index == 0:
values[0] = float(x)
else:
values[1][str(index)+"_"+x] = 1
return values
n_feature = 2**14
hasher = FeatureHasher(n_features=n_feature)
training_file_hashed = training_file.map(lambda line: [hash_feature(line)[0], hasher.transform([hash_feature(line)[1]])])
def build_lable_points(line):
values = [0.0] * n_feature
for index, value in zip(line[1].indices, line[1].data):
values[index] = value
return LabeledPoint(line[0], values)
parsed_training_data = training_file_hashed.map(lambda line: build_lable_points(line))
model = LogisticRegressionWithSGD.train(parsed_training_data)
The error happens when executing the last line.
Answer
The Integer.MAX_INT restriction is on the size of a file being stored. 1.2M rows is not a big thing, to I'm not sure your problem is "the limits of spark". More likely, some part of your work is creating something too big to be handled by any given executor.
I'm no Python coder, but when you "hashed the features of the records" you might be taking a very sparse set of records for a sample and creating an non-sparse array. This will mean a lot of memory for 16384 features. Particularly, when you do zip(line[1].indices, line[1].data). The only reason that doesn't get you out of memory right there is the shitload of it you seem to have configured (50G).
Another thing that might help is to increase the partitioning. So if you can't make your rows use less memory, at least you can try having fewer rows on any given task. Any temporary files being created are likely to depend on this, so you'll be more unlikely to hit file limits.
And, totally unrelated to the error but relevant for what you are trying to do:
16384 is indeed a big number of features, in the optimistic case where each one is just a boolean feature, you have a total of 2^16384 possible permutations to learn from, this is a huge number(try it here: https://defuse.ca/big-number-calculator.htm).
It is VERY, VERY likely that no algorithm will be able to learn a decision boundary with just 1.2M samples, you would probably need at least a few trillion trillion examples to make a dent on such a feature space. Machine Learning has its limitations, so don't be surprised if you don't get better-than-random accuracy.
I would definitely recommend trying some sort of dimensionality reduction first!!