如何在Hortonworks VM上的Spark上运行.jar？

Question

I'm new to Hortonworks VM and I got confused. I'm trying to run a .jar file on Spark. Normally I test locally on Windows by running

spark-submit --driver-memory 4g --class en.name.ClassName %CODE%/target/program.jar

but since I need Hive, I thought I'd move to a Hortonworks VM to test locally. Now, I've uploaded my .jar and the input files to the HDFS (to the /tmp/my_code directory) via Hortonworks' Ambari's HDFS Files GUI. What next? I also found the command line, but how do I access my .jar on the HDFS from the VM's command line? I'm trying to run

spark-submit --driver-memory 4g --class en.name.ClassName /tmp/my_code/program.jar from the sandbox condole (the one running on http://127.0.0.1:4200/ by default, root@sandbox "Shell in a Box"), which is not working. It says the .jar does not exists. How can I point the VM to use the .jar on HDFS? Thank you!

Answer 1

JAR should be on local file system NOT in hdfs . Only input file should be in hdfs . so /tmp/my_code/program.jar path should be local, and that why you are seeing that error: the .jar does not exists

if you run-this command:

>spark-submit --help

you will see:

 --jars JARS   Comma-separated list of local jars to include on the driver
               and executor classpaths.

update: Accroding to Documentations :

application-jar: Path to a bundled jar including your application and all dependencies. The URL must be globally visible inside of your cluster, for instance, an hdfs:// path or a file:// path that is present on all nodes.

so,

if jar is on hdfs:

spark-submit --driver-memory 4g --class en.name.ClassName hdfs://target/program.jar

if jar is on local:

spark-submit --driver-memory 4g --class en.name.ClassName /target/program.jar

OR

spark-submit --driver-memory 4g --class en.name.ClassName file://target/program.jar