This issue is best explained with an example:

There is a github organization called 'infra' with two repos:

• infra/testrepo
• infra/testrepo2

These 2 repos are indexed and running in the Github organization job. 

There is also a normal pipeline job (pipeline-test) created that has 'Pipeline script from SCM' configured, pulling the Jenkinsfile from infra/testrepo2 (so it use the exact same configuration)

properties([pipelineTriggers([upstream(threshold: hudson.model.Result.SUCCESS, upstreamProjects: 'infra/testrepo/master')])])

node('mynode'){
  echo "foo"
}

Expected: When 'infra/testrepo/master' is ran, either manually or with SCM trigger in the Github Organization Job, this should trigger 'pipeline-test' and 'infra/testrepo2/master' because of the pipelineTrigger set in the Jenkinsfile above.

**Note: testrepo2 is built at least once, to pull in the properties and have the trigger be set up, and I see this reflected when clicking 'View Configuration'. (attached build_triggers.jpeg shows that the correct build trigger is present in both pipeline-test and infra/testrepo2/master so the properties did get read in)

Actual: Running 'infra/testrepo/master' only triggers 'pipeline-test' and not 'infra/testrepo2/master'

in IRC Jesse Glick said this may be due to a missed call to 'Trigger.start' somewhere