How to pass arguments in a java action within an oozie workflow?

Question

I'm trying to create a Oozie workflow with one action, a java action to launch a program to parse an excel file inside HDFS, so, in order to this java program parse the file I need to pass it the location of the file, without the oozie workflow I just call the jar file with the arguments I need like:

   java -jar myJarFile.jar hdfs:///path/to/the/file.xlsx

Now in the java action inside the oozie workflow I need to write something like:

<action name="parse-file">
 <java>
  <job-tracker>${jobTracker}</job-tracker>
  <name-node>${nameNode}</name-node>
  <main-class>parseFile</main-class>
  <arg>path/to.the/file.xlsx</arg>
  <file>lib/myJarFile.jar#myJarFile.jar</file>
 </java>
 <ok to="sen-success-email"/>
 <error to="failed-notification-email"/>
</action>

Is this approach correct?

attila_s · Accepted Answer · 2019-06-03 23:29:35Z

1

The java action will execute the public static void main(String[] args) of your class. Make sure you use fully specified package name. For example a.b.c.MainClass
You don't need the <file> tag if you store myJarFile.jar in a lib directory next your workflow.xml on HDFS. For example:

workflow_dir/workflow.xml
workflow_dir/lib
workflow_dir/lib/myJarFile.jar

Make sure you also upload all dependencies with your workflow.

References:

Description of the Java action: https://oozie.apache.org/docs/5.1.0/WorkflowFunctionalSpec.html#a3.2.6_Java_Action
Sample workflow xml for a Java action: https://github.com/apache/oozie/blob/master/examples/src/main/apps/java-main/workflow.xml#L30

answered Jun 3, 2019 at 23:29

attila_s

1709 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

EEstereo Over a year ago

All right, this is helpful, just one thing, for example I'm running this jar file from the command line:

java -cp FileProcessosSQL.jar:'hadoop classpath':'hdfs classpath':'hbase classpath' com.kata.team1.ProcessSQL SqlInsertion.sql

Just to be clear about it, in my folder workflow/lib I should put the libraries that are in hadoop class path, HDFS class path and hbase class path ?

attila_s Over a year ago

That should work. Alternatively, you could also upload dependencies to HDFS, and use oozie.libpath=/path/to/jars,another/path/to/jars in job.properties or define your own sharelib.

Collectives™ on Stack Overflow

How to pass arguments in a java action within an oozie workflow?

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related