Spark running Liblinear unable to load JBLAS jar - scala

I'm running spark 1.4.0, hadoop 2.7.0, and JDK 7. I'm trying to run the example code of Liblinear presented here.
The liblinear jar works, however when training the model it can't find the JBLAS library. I've tried including a JBLAS library in the --jars option when launching spark, as well as installing the jar with maven (although I must add I am a newbie to spark as well as maven so I probably did it wrong).
The specific error thrown is this:
java.lang.NoClassDefFoundError: org/jblas/DoubleMatrix
at tw.edu.ntu.csie.liblinear.Tron.tron(Tron.scala:323)
at tw.edu.ntu.csie.liblinear.SparkLiblinear$.tw$edu$ntu$csie$liblinear$SparkLiblinear$$train_one(SparkLiblinear.scala:32)`
when running this line:
val model = SparkLiblinear.train(data, "-s 0 -c 1.0 -e 1e-2")`
Thanks.

java.lang.NoClassDefFoundError: org/jblas/DoubleMatrix
It seems that you did not add jblas jar. The solution could be:
$ export SPARK_CLASSPATH=$SPARK_CLASSPATH:/path/to/jblas-1.2.3.jar
After that, it would work fine.
Hope this helps,
Le Quoc Do

Related

ClassNotFound Exception in Spark Scala Code

I have used maven project in Scala. I have used all dependencies in pom.
Still I am getting ClassNotFoundException when I run spark-submit command.
clean compile assembly:single is the Maven goal which I used.
Following is the spark submit command I have used.
spark-submit --class com.SimpleScalaApp.SimpleApp --master local /location/file.jar
Unfortunately there is no useful comment on the question in my opinion. The ClassNotFoundException mostly is thrown when version of your dependencies are not match together.
First of all it's better to check if all your dependencies are matched by checking those dependencies' pom files and consequently their usage order in the main project pom file.
A short description about the best practice of running the program on spark cluster is using shaded or assembly build form.

How to compile and build spark examples into jar?

So I am editing MovieLensALS.scala and I want to just recompile the examples jar with my modified MovieLensALS.scala.
I used build/mvn -pl :spark-examples_2.10 compile followed by build/mvn -pl :spark-examples_2.10 package which finish normally. I have SPARK_PREPEND_CLASSES=1 set.
But when I re-run MovieLensALS using bin/spark-submit --class org.apache.spark.examples.mllib.MovieLensALS examples/target/scala-2.10/spark-examples-1.4.0-hadoop2.4.0.jar --rank 5 --numIterations 20 --lambda 1.0 --kryo data/mllib/sample_movielens_data.txt I get java.lang.StackOverflowError even though all I added to MovieLensALS.scala is a println saying that this is the modified file, with no other modifications whatsoever.
My scala version is 2.11.8 and spark version is 1.4.0 and I am following the discussion on this thread to do what I am doing.
Help will be appreciated.
So I ended up figuring it out myself. I compiled using mvn compile -rf :spark-examples_2.10 followed by mvn package -rf :spark-examples_2.10 to generate the .jar file. Note that the jar file produced here is spark-examples-1.4.0-hadoop2.2.0.jar.
On the other hand, the stackoverflow error was because of a long lineage. For that I could either use checkpoints of reduce numiterations, I did the later. I followed this for it.

Building Customize Spark

We are creating a customize version of Spark since we are changing some lines of code from ALS.scala. We build the customize spark version using
mvn command:
./make-distribution.sh --name custom-spark --tgz -Psparkr -Phadoop-2.6 -Phive -Phive-thriftserver -Pyarn
However, upon using the customized version of Spark, we run into this error:
Do you guys have some idea on what causes the error and how we might solve the issue?
I am actually using a jar file in the local machine by building them using sbt: sbt compile then sbt clean package and putting the jar file here: /Users/user/local/kernel/kernel-0.1.5-SNAPSHOT/lib.
However in the hadoop environment, the installation is different. Thus, I use maven to build spark and that's where the error comes in. I am thinking that this error might be dependent on using maven to build spark as there are some reports like this:
https://issues.apache.org/jira/browse/SPARK-2075
or probably on building spark assembly files

Exception in thread "main" java.lang.NoClassDefFoundError: org/apache/spark/rdd/RDD

Please note that I am better dataminer than programmer.
I am trying to run examples from book "Advanced analytics with Spark" from author Sandy Ryza (these code examples can be downloaded from "https://github.com/sryza/aas"),
and I run into following problem.
When I open this project in Intelij Idea and try to run it, I get error "Exception in thread "main" java.lang.NoClassDefFoundError: org/apache/spark/rdd/RDD"
Does anyone know how to solve this issue ?
Does this mean i am using wrong version of spark ?
First when I tried to run this code, I got error "Exception in thread "main" java.lang.NoClassDefFoundError: scala/product", but I solved it by setting scala-lib to compile in maven.
I use Maven 3.3.9, Java 1.7.0_79 and scala 2.11.7 , spark 1.6.1. I tried both Intelij Idea 14 and 15 different versions of java (1.7), scala (2.10) and spark, but to no success.
I am also using windows 7.
My SPARK_HOME and Path variables are set, and i can execute spark-shell from command line.
The examples in this book will show a --master argument to sparkshell, but you will need to specify arguments as appropriate for your environment. If you don’t have Hadoop installed you need to start the spark-shell locally. To execute the sample you can simply pass paths to local file reference (file:///), rather than a HDFS reference (hdfs://)
The author suggest an hybrid development approach:
Keep the frontier of development in the REPL, and, as pieces of code
harden, move them over into a compiled library.
Hence the samples code are considered as compiled libraries rather than standalone application. You can make the compiled JAR available to spark-shell by passing it to the --jars property, while maven is used for compiling and managing dependencies.
In the book the author describes how the simplesparkproject can be executed:
use maven to compile and package the project
cd simplesparkproject/
mvn package
start the spark-shell with the jar dependencies
spark-shell --master local[2] --driver-memory 2g --jars ../simplesparkproject-0.0.1.jar ../README.md
Then you can access you object within the spark-shell as follows:
val myApp = com.cloudera.datascience.MyApp
However if you want to execute the sample code as Standalone application and execute it within idea you need to modify the pom.xml.
Some of dependencies are required for compilation, but are available in an spark runtime environment. Therefore these dependencies are marked with scope provided in the pom.xml.
<!--<scope>provided</scope>-->
you can remake the provided scope, than you will be able to run the samples within idea. But you can not provide this jar as dependency for the spark shell anymore.
Note: using maven 3.0.5 and Java 7+. I had problems with maven 3.3.X version with the plugin versions.

Why does running Spark job fail to find classes inside uberjar on EMR while it works locally fine?

I have a Spark Job that is using some external libraries to work. When I run the job locally through the main method from IntelliJ the job runs without any issues. However, when I assembly my job into a jarfile (I create an UberJAR using sbt) and I try to run it on EMR, it throws a ClassNotFoundException.
I have checked that the class is indeed inside the jarfile so it should be available for the job to run. I have also tried the spark-submit options spark.driver.extraClassPath, spark.driver.extraLibraryPath, spark.executor.extraClassPath and spark.executor.extraLibraryPath as well as spark.driver.userClassPathFirst and spark.executor.userClassPathFirst. Also, I tried doing in the code sparkContext.addJar("/mnt/jars/myJar"). None of them worked for me.
Also, when running on EMR I can read the log that says that the JAR was added (not sure if it is loaded on the classpath, but it should because other classes are being loaded properly):
15/11/02 04:10:26 INFO SparkContext: Added JAR file:///mnt/my-app-1.0-SNAPSHOT.jar at http://172.31.42.244:44471/jars/my-app-1.0-SNAPSHOT.jar with timestamp 1446437426661
I am running out of ideas about what else to try. I have been researching and I see few tickets on the Spark JIRA board but nothing similar to my issue.
I am running on EMR release-label 4.1.0 (Spark 1.5.0), Java 7, sbt 0.13.7 and Scala 2.10.5.
I think when launching your job on EMR you need to provide the s3 location for your jar dependencies a la the manual e.g. -u s3://sparksupport/libs. These jars will be added to the classpath when running spark.
It turned out to be a problem with SerializationUtils from Apache Commons Lang. There is an open issue where the class will throw a ClassNotFoundException even if the class is in the classpath in a multiple-classloader environment: https://issues.apache.org/jira/browse/LANG-1049
We moved away from the library and our Spark job is working fine now. The issue was not related with Spark finally.