Hadoop: how to remote debug single oozie java workflow

Question

I have been unable to connect a debugger to a java workflow .

I do not wish to open up the whole of the hadoop jobs to debug (as specified in many places, eg How to debug hadoop mapreduce jobs from eclipse? , or REMOTE DEBUGGING OF HADOOP JOB WITH ECLIPSE , etc...). I wish to debug a single(!) workflow.

I have a small cluster (3 machines), and I wish to debug in this simulated dev environment .

I had tried to specify the following in the workflow:

<property>
    <name>mapred.child.java.opts</name>
    <value>-agentlib:jdwp=transport=dt_socket,server=y,suspend=y,address=12121</value>
</property>

But this had not worked.

One thing else is that I am actually running this as a sub-workflow

Suggestions?

Answer 1

You only have to include this ...

<property>
   <name>oozie.launcher.mapred.child.java.opts</name>
   <value>-agentlib:jdwp=transport=dt_socket,server=y,suspend=y,address=12121</value>
</property>     

or, in my case ...

<property>
   <name>oozie.launcher.mapred.child.java.opts</name>
   <value>-Xdebug -Xrunjdwp:transport=dt_socket,server=y,suspend=y,address=7777</value>
</property>