[jira] [Resolved] (NIFI-9193) Generateflowfile must be able to run on a single node

["Dennis Jaheruddin (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/NIFI-9193?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Dennis Jaheruddin resolved NIFI-9193.
-------------------------------------
    Resolution: Invalid

I got confused after seeing that the GenerateFlowFile processor does not report the data it generates as ingestion, however the actual validation is done in a completely different way (looking at inbound connections). 

 

Closing this issue.

> Generateflowfile must be able to run on a single node
> -----------------------------------------------------
>
>                 Key: NIFI-9193
>                 URL: https://issues.apache.org/jira/browse/NIFI-9193
>             Project: Apache NiFi
>          Issue Type: Improvement
>          Components: Core Framework
>    Affects Versions: 1.14.0
>            Reporter: Dennis Jaheruddin
>            Priority: Major
>
> If I understand the release notes of 1.14 correctly the option has been removed to set processors without incoming connections to run on the primary node.
> I understand that the primary node should be used for lightweight things that must only be done once, like listing, and therefore that there should hardly be a reason to allow other processors to run as such.
> However, I believe the GenerateFlowFile processor is a special case, its main purpose is testing or simulating, and it is very convenient when there is no outside connection available, to develop or test something. 
> As such, the GenerateFlowFile processor should be allowed to run on the primary node, so it may substitute any processor (or set of processors) during development.
> Alternately (or in addition) it may make sense to at least allow this to run on a single node, but for most consistency it would be good to allow for Primary node running. 
>  
>  
> ----
> There may be ways to achieve a similar result (e.g. routing its output to single node and then throttling) but this is quite awkward.
> This issue is mostly about being able what we were able to do before. However, there is an additional benefit in that now autoscaling is becoming more dominant, and it is also very convenient to test autoscaling if you can predictably generate an amount of flowfiles (withouth having the generator scale up when the nodes scale up).



--
This message was sent by Atlassian Jira
(v8.3.4#803005)