Glad to hear there is a root cause for this. Am I correct in assuming that I won't be able to just rename the offending input port to get this current change pushed out? Basically the "bad thing" is already embedded in the change history of the flow?
BI CI Solutions Architect
From: Kevin Doran [mailto:[EMAIL PROTECTED]]
Sent: Thursday, January 10, 2019 4:41 PM
To: [EMAIL PROTECTED]; Russ Weiher <[EMAIL PROTECTED]>
Subject: RE: Java exception when attempting to update version from registry
The diff was perfect. It showed the problem was two input ports with the same name but different IDs, which I was able to reproduce. Your stack trace and flow diff helped us uncover a bug in our update flow logic -- thanks!
I verified it still affects the latest version of NiFi and I opened a Jira issue  that captures this. Feel free to sign up to ASF Jira to "watch" that issue if you want to see when this gets patched. In the meantime, you can avoid this by avoiding deleting/adding input ports with the same name (during a single version change) within a versioned process group. It's a pretty rare case that should not come up often, so I don't think this should impact your workflow for moving flows from dev to higher envs in general. If you do run into other failures in your workflow, please let us know.
On January 10, 2019 at 15:45:43, Russ Weiher ([EMAIL PROTECTED]) wrote:
IMPORTANT: This email (and any attachments) is intended for the use of only the person or entity to which it is addressed, and may contain information that is privileged and confidential. You, the recipient, are obligated to maintain it in a safe, secure and confidential manner. Unauthorized redisclosure or failure to maintain confidentiality may subject you to federal and state penalties. If you are not the intended recipient, please immediately notify us by return email, and delete this message from your computer.