withMaven dependency trigger - Need help to understand the error message

[Dan Tran]

Hi 

I am trying to get an upstream to trigger downstream project using snapshot trigger. My upstream logs the following message

   [withMaven] Not triggering xxxxxx because it has dependencies in the downstream project list

Could some one explain what it really mean?

'xxxxxxx'  is my downstream project, looks like I have he right setup where a maven module of a downstream project depends on other maven module to the upstream project

Thanks

-Dan

[Cyrille Le Clerc]

Sample: projects framework, shared-lib and webapp.

Both shared-lib and webapp depend on framework and webapp also depend on shared-lib.

When framework build succeeds, it should trigger both shared-lib and webapp. But it's a "waste" to trigger webapp now because shared-lib will trigger "webapp" when it succeeds.

This is the message you get.

[Dan Tran]

Thanks for the explanation.  It is clear now

In my case, I downstream should be built regardless, but it is not.  

How can I troubleshoot this?

Thanks

-D

[Cyrille Le Clerc]

> How can I troubleshoot this?

I discussed this feature with Arnaud Héritier today for another use case.

My understanding is that it is "work as design" and that you ask for an improvement to disable this JENKINS-46313 Avoid excessive triggering of downstream jobs capability.

Do you have the same understanding of the status? If so, can you please open an RFE ticket on Jira?

Cyrille

https://github.com/jenkinsci/pipeline-maven-plugin/blob/pipeline-maven-3.4.3/jenkins-plugin/src/main/java/org/jenkinsci/plugins/pipeline/maven/listeners/DownstreamPipelineTriggerRunListener.java#L74

[Dan Tran]

I believe I don't have the scenario that you described.  Mine is more like    Project A, B, and C   depends on D,   after D builds, A, B, and C are not built

So I need to troubleshoot, to see what I missed from my setup

-D

[Cyrille Le Clerc]

Ack.

We need to improve the troubleshooting guide "Why my downstream job doesn't get triggered".

A SQL approach could be interesting but as our queries are no trivial, there would be a risk of misunderstanding. Maybe adding a troubleshooting API/GUI/CLI on the withMaven pipeline graph is the most efficient way.

Cyrille