Github user mattyb149 commented on the issue:
https://github.com/apache/nifi/pull/2773
@markap14 How would we get at the value in NiFiProperties? They're not
readily available from the FlowController (nor is the FlowController interface
itself IIRC). Perhaps we call a static method
Github user markap14 commented on the issue:
https://github.com/apache/nifi/pull/2773
@mattyb149 I think this makes sense but have a couple of thoughts on the PR:
- Probably should use the value set in NiFiProperties for the http/https
hostname if it is populated. This would be
Github user mcgilman commented on the issue:
https://github.com/apache/nifi/pull/2773
@zenfenan Correct. The cluster coordinator handles setting the node address
for the bulletins that are returned through the REST API. However, those are
not the same objects that are available in
Github user zenfenan commented on the issue:
https://github.com/apache/nifi/pull/2773
@mcgilman Are you saying regardless of the `setNodeAddress` API not setting
anything, the cluster coordinator takes care of populating it ?
---
Github user mcgilman commented on the issue:
https://github.com/apache/nifi/pull/2773
I'm not exactly sure why the Bulletin itself has a `nodeAddress` field.
However, I can offer that the corresponding BulletinDTO contains a
`nodeAddress` to help differentiate between which node in
Github user zenfenan commented on the issue:
https://github.com/apache/nifi/pull/2773
I'm not entirely sure if this poses an immediate threat, but in the future
when other extensions try to use the `getNodeAddress()` API, it might possibly
be an issue, especially on the user
Github user mattyb149 commented on the issue:
https://github.com/apache/nifi/pull/2773
@zenfenan Good point. The reporting tasks haven't had this information
available before, is this a "no-harm-no-foul" thing, and if not, how should it
be handled?
---