[jira] [Commented] (NIFI-333) Some Processors catch java.land.Exception and route to 'failure', instead of catching ProcessException

["Dan Bress (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/NIFI-333?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14321586#comment-14321586 ] 

Dan Bress commented on NIFI-333:
--------------------------------

Working on a fix for this now.  Saw cases of this for the following processors:
* CreateHadoopSequenceFile
* Base64EncodeContent
* CompressContent
* HashContent
* PutEmail
* TransformXml

Also noticed Exception being caught in SegmentContent, which is then throwing ProcessException.  Guessing thats OK, it doesn't have a failure relationship.

Noticed a few processors catching Throwable, which I didn't touch:
* GetHTTP
* PutFile


> Some Processors catch java.land.Exception and route to 'failure', instead of catching ProcessException
> ------------------------------------------------------------------------------------------------------
>
>                 Key: NIFI-333
>                 URL: https://issues.apache.org/jira/browse/NIFI-333
>             Project: Apache NiFi
>          Issue Type: Bug
>            Reporter: Daniel Ueberfluss
>
> The consequence of this is that if the framework throws a FlowFileAccessException, for example, the Processor is catching that and then routing to failure, when the framework should be able to catch it and roll back the ProcessSession.  Specifically, we are seeing this because on of the content repository containers ran out of disk space and files were then routed to 'failure'. Instead, the session should have been rolled back and the FlowFile processed again. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)