spark-core 1.6.1 & lift-json 2.6.3 java.lang.NoClassDefFoundError

问题

I have a Spark application which has a sbt file just like below.
It works on my local machine. But when I submit it to EMR running Spark 1.6.1, an error occured like below:

java.lang.NoClassDefFoundError: net/liftweb/json/JsonAST$JValue

I am using "sbt-package" to get jar

Build.sbt:

organization := "com.foo"
name := "FooReport"

version := "1.0"

scalaVersion := "2.10.6"

libraryDependencies ++= Seq(
  "org.apache.spark" %% "spark-core" % "1.6.1"
  ,"net.liftweb" % "lift-json_2.10" % "2.6.3"
  ,"joda-time" % "joda-time" % "2.9.4"
)

Do you have any idea about what’s happening?

回答1:

I' ve found a solution and it is working!

The problem was all about sbt package which doesn't include all dependent jars to output jar. To overcome this I tried sbt-assembly but I got plenty of "deduplicate" errors when I ran it.

After all I came up to this blog post which made everthing clear.
http://queirozf.com/entries/creating-scala-fat-jars-for-spark-on-sbt-with-sbt-assembly-plugin

In order to submit Spark jobs to a Spark Cluster (via spark-submit), you need to include all dependencies (other than Spark itself) in the Jar, otherwise you won't be able to use those in your job.

Create "assembly.sbt" under /project folder.
Add this line addSbtPlugin("com.eed3si9n" % "sbt-assembly" % "0.14.3")
Then paste the assemblyMergeStrategy code below to your build.sbt

assemblyMergeStrategy in assembly := { case PathList("javax", "servlet", xs @ _*) => MergeStrategy.last case PathList("javax", "activation", xs @ _*) => MergeStrategy.last case PathList("org", "apache", xs @ _*) => MergeStrategy.last case PathList("com", "google", xs @ _*) => MergeStrategy.last case PathList("com", "esotericsoftware", xs @ _*) => MergeStrategy.last case PathList("com", "codahale", xs @ _*) => MergeStrategy.last case PathList("com", "yammer", xs @ _*) => MergeStrategy.last case "about.html" => MergeStrategy.rename case "META-INF/ECLIPSEF.RSA" => MergeStrategy.last case "META-INF/mailcap" => MergeStrategy.last case "META-INF/mimetypes.default" => MergeStrategy.last case "plugin.properties" => MergeStrategy.last case "log4j.properties" => MergeStrategy.last case x => val oldStrategy = (assemblyMergeStrategy in assembly).value oldStrategy(x) }

And run sbt assembly

Now you have a big fat jar which has all the dependencies. It might be hundreds of MB based on dependent libraries. For my case I am using Aws EMR which Spark 1.6.1 is already installed on it. To exclude spark-core lib from your jar you can use "provided" keyword:

"org.apache.spark" %% "spark-core" % "1.6.1" % "provided"

Here is the final build.sbt file:

organization := "com.foo"
name := "FooReport"

version := "1.0"

scalaVersion := "2.10.6"

libraryDependencies ++= Seq(
  "org.apache.spark" %% "spark-core" % "1.6.1" % "provided"
  ,"net.liftweb" % "lift-json_2.10" % "2.6.3"
  ,"joda-time" % "joda-time" % "2.9.4"
)

assemblyMergeStrategy in assembly := {
  case PathList("javax", "servlet", xs @ _*) => MergeStrategy.last
  case PathList("javax", "activation", xs @ _*) => MergeStrategy.last
  case PathList("org", "apache", xs @ _*) => MergeStrategy.last
  case PathList("com", "google", xs @ _*) => MergeStrategy.last
  case PathList("com", "esotericsoftware", xs @ _*) => MergeStrategy.last
  case PathList("com", "codahale", xs @ _*) => MergeStrategy.last
  case PathList("com", "yammer", xs @ _*) => MergeStrategy.last
  case "about.html" => MergeStrategy.rename
  case "META-INF/ECLIPSEF.RSA" => MergeStrategy.last
  case "META-INF/mailcap" => MergeStrategy.last
  case "META-INF/mimetypes.default" => MergeStrategy.last
  case "plugin.properties" => MergeStrategy.last
  case "log4j.properties" => MergeStrategy.last
  case x =>
    val oldStrategy = (assemblyMergeStrategy in assembly).value
    oldStrategy(x)
}

来源：https://stackoverflow.com/questions/38335207/spark-core-1-6-1-lift-json-2-6-3-java-lang-noclassdeffounderror

标签

scala

apache-spark

sbt

lift-json