[ 
https://issues.apache.org/jira/browse/NIFI-11570?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17799130#comment-17799130
 ] 

Nissim Shiman commented on NIFI-11570:
--------------------------------------

Circling back to this...

This will involve changes to flow.json so I am hoping to have more people 
comment before comitting to an approach.

Thank you very much [~bbende]  for your approach (linked above) of
storing the set of sensitiveDynamicPropertyNames in flow.json.

I ran this problem by a senior engineer, who suggested that using the 
(currently existing, but empty) propertyDescriptor object in flow.json is a 
good approach as well.

This may end of touching many files, so I would very much appreciate more 
feedback on this (more approaches or comments to the above two ideas)
 

> Processors have difficulty restoring dynamic properties when leaving ghost 
> mode
> -------------------------------------------------------------------------------
>
>                 Key: NIFI-11570
>                 URL: https://issues.apache.org/jira/browse/NIFI-11570
>             Project: Apache NiFi
>          Issue Type: Bug
>          Components: Core Framework
>    Affects Versions: 1.19.1
>            Reporter: Nissim Shiman
>            Assignee: Nissim Shiman
>            Priority: Major
>
> When a processor with non sensitive dynamic properties is started in ghost 
> mode due to its underlying nar being missing, it will have a difficult time 
> being restored even when its nar has been replaced.
> The dynamic properties will be identified as being sensitive and not able to 
> be accessed.
> So for example, if a dynamic property added to GenerateFlowFile and 
> nifi-standard-nar is removed from lib and nifi is restarted, GenerateFlowFile 
> will come up in ghosted mode (as expected).
> When nifi-standard-nar is replaced and nifi is restarted, GenerateFlowFile 
> will be restored, but its dynamic property will not be accessable.
> This was originally filed as a second part of NIFI-11109, but after looking 
> more into the underlying causes/symptoms, realized that these are two 
> different issues. 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to