I had a conflict with a rapid miner jar with the version of the java poi library that I was using in my spark submit and I saw that there are some options in oozie to try to give my jar precedence over other jars on the hadoop class path / bundled with oozie.
Attached is my oozie workflow https://paste.ofcode.org/fdhZTXwQ9FwNwUMv4LevWJ .
The problem is after adding these configurations my oozie workflow just hangs and doesnt even start.
Similiar problem here https://grokbase.com/t/cloudera/cdh-user/142kqbnk21/classpath-precedence-in-oozie
Tried adding these lines but now the workflow just hangs.
<configuration> <property> <name>oozie.launcher.mapreduce.task.classpath.user.precedence</name> <value>true</value> </property> <property> <name>oozie.launcher.mapreduce.job.classpath.user.precedence</name> <value>true</value> </property> </configuration>
Guess these configs should be set inside the coordinator.xml as well, take a look at the example below
And also, I found a link that might help you with your jar priority over hadoop classpath.
Hope this helps!