All I cared about was getting livy-client-http to build and installed in my local maven .m2 repository. This is what I did in Windows 7:
git clone
incubator-livy then comment out this section in its pom.xml
(credit to Yatzhash):
<requireOS>
<family>unix</family>
</requireOS>
Run mvn install -DskipTests
in the top level directory. If you do this from a regular Windows command prompt it will eventually fail with this error:
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-antrun-plugin:1.8:run (default) on project livy-server: An Ant BuildException has occured: Execute
failed: java.io.IOException: Cannot run program "bash" (in directory "C:\github.com\incubator-livy\server"): CreateProcess error=2, The system cannot find the
file specified
[ERROR] around Ant part ...... @ 4:27 in C:\github.com\incubator-livy\server\target\antrun\build-main.xml
But if you run the same command in a Git Bash prompt instead you can bypass this error. Eventually the build will fail at livy-integration-test but at least the livy-client-http build should have passed.
But you'll find this jar will have been installed in your .m2 repository: livy-client-http-0.4.0-incubating-SNAPSHOT.jar. This means you need to modify your own client app's dependency to look like this instead of the one recommended by the Livy docs:
<dependency>
<groupId>org.apache.livy</groupId>
<artifactId>livy-client-http</artifactId>
<version>0.4.0-incubating-SNAPSHOT</version>
</dependency>
I also had to add this dependency to my client app because I wanted to talk to a Spark 2.1 server via Livy:
<dependency>
<!-- See https://spark.apache.org/docs/2.1.0/programming-guide.html#linking-with-spark -->
<groupId>org.apache.spark</groupId>
<artifactId>spark-core_2.11</artifactId>
<version>2.1.0</version>
</dependency>
Then my client app would compile in Windows.