I have been working with Oozie for quite a few weeks, and the experience so far has been quite positive. It is quite easy to learn, given that you understands XML and Hadoop ecosystem. However, there is one limitation that is quite annoying, although I understand the purpose behind the design. When you have a decision making node that branches to another node in the later stage, you can not have other nodes go to that particular node anymore in the same XML workflow. To illustrate the problem, see the workflow XML below:

    
        ${ wf:conf('my_variable') == 'true'}
        
    


    
    


    
    

You can see that “node2” can be reached either from “decision” node, or from “node1” node. This is not allowed in Oozie’s workflow by default, and you will get the following error:
Error: E0743 : E0743: Multiple "ok to" transitions to the same node are not allowed
In order to force it to run, you will need to set the following config in the job properties file:
oozie.validate.ForkJoin=false
This should get Oozie goind without complaining the potential problems.

Leave a Reply

Your email address will not be published. Required fields are marked *