Santiago,

OK great, i’m glad you were able to get it working. So I think the issue is 
that the script, when the getRelationships() method is called, should return a 
Set<Relationship> but appears to return a Set<ArrayList> - probably a 
Set<ArrayList<Relationship>>.
But the Objects that are in that Set are not of type Relationship - they are of 
type ArrayList.

So that’s causing Validation to fail. I think the validation is unrelated, but 
that same problem was causing failures elsewhere also.

So when you disabled and restarted, the processor will have not parsed the 
script so it’s only returning the relationships it knows about, so it’s working.

Hope this helps to clarify what’s happening!

Thanks
-Mark




On Mar 11, 2021, at 1:55 PM, Santiago Acosta 
<[email protected]<mailto:[email protected]>>
 wrote:

I was successfull in recreating my experimental setup and the issue. With what 
you told me about disabled processors, I understand know that a service restart 
would be unhindered by the misbehaving processor. The restart was successful 
and the restart on my original experimental setup was also successful. I could 
then access the PG and delete the InvokeScriptedProcessor.

If anyone ever comes across this issue, keep in mind that it may have only 
worked because I was able to disable the entire PG (including the misbehaving 
one). I believe that the error only propagated to the point where my Parameter 
Context was defined as this processor was nested 2 or 3 levels deep. A diagram 
will help:

NiFi Flow [still accessible]
  / PG (that defines a PContext1) [cannot access]
    / PG 1 (with PC1)
      / PG 2 (with PC1)
        / InvokeScriptedProcessor (misbehaving)

I could still access any other PG from the root PG. I don't know if this is 
relevant but it was part of my setup.

Thank you Mark.

--
Best regards,
Santiago

Reply via email to