Configuring the Oozie Workflow Activity stage (Linux®)

If you installed IBM® InfoSphere® BigInsights, you can use the Oozie Workflow Activity stage to invoke Oozie jobs from the Designer client. To use the Oozie Workflow Activity stage, you must extract the files that the Oozie client API requires to run the invoker utility.

About this task

The Oozie client API is installed when you install InfoSphere BigInsights. You must ensure that InfoSphere DataStage® can access the following .jar files that the Oozie client API requires. The names of the .jar files might differ, depending on the version of Oozie that you installed.
  • oozie-client-2.3.1.jar
  • json-simple-1.1.jar
  • commons-cli-1.2.jar

This procedure uses InfoSphere BigInsights as an example, but the procedure applies to any Hadoop installation. On other systems, such as Cloudera, you must configure Oozie before you can configure the Oozie Workflow Activity stage.

Procedure

  1. Copy the Oozie client API files.
    1. On the machine that is running InfoSphere BigInsights, open a shell prompt and log in as the root user.
    2. Navigate to the $BIGINSIGHTS_HOME/oozie directory, where $BIGINSIGHTS_HOME is the directory where you installed InfoSphere BigInsights.
    3. Copy the Oozie client API files to the machine that is running the InfoSphere DataStage engine. The files are contained in compressed file like oozie-client-2.3.1.tar.gz. The file name might differ, depending on the version of Oozie that you installed.
  2. Make the Oozie client API .jar file available to the InfoSphere DataStage engine.
    1. On the machine that is running the InfoSphere DataStage engine, log in as the InfoSphere DataStage administrator user.
    2. Extract the .jar files from the compressed Oozie client API file to a directory that is accessible by the InfoSphere DataStage administrator user. For example, /home/dsadm/oozie.
    3. Run the chmod command to grant InfoSphere DataStage access to the .jar files. 
      chmod o+x *
    4. Navigate to the opt/IBM/InformationServer/Server/DSEngine directory and open the dsenv file in an editor.
    5. Add the following lines to the end of the dsenv file. 
      # Oozie client library location
OOZ_PATH=/home/dsadm/oozie

# Dependent Oozie .jar files
JAR1=$OOZ_PATH/oozie-client-2.3.1.jar
JAR2=$OOZ_PATH/json-simple-1.1.jar
JAR3=$OOZ_PATH/commons-cli-1.2.jar

CLASSPATH=${JAR1}:${JAR2}:${JAR3}:${CLASSPATH}

      OOZ_PATH is the directory where you extracted the .jar files from the compressed Oozie client API file. The file names might differ, depending on the version of Oozie that you installed.

    6. Save and close the dsenv file.
  3. Restart the InfoSphere DataStage engine.
    1. Navigate to the opt/IBM/InformationServer/Server/DSEngine directory.
    2. Run the following command to start the InfoSphere DataStage engine. 
      bin/uv -admin -stop
    3. Run the following command to restart the InfoSphere DataStage engine. 
      bin/uv -admin -start