[FIXED] Jenkins scripted pipeline doesn't return all causes for parallel build job failures

Issue

I have a simple scripted pipeline as follows -

import hudson.model.*
import groovy.json.*

def map = [:]
map['second'] = {
    build job: "childPipeline1", propagate: true, wait: true
}

map['third'] = {
    build job: "childPipeline2", propagate: true, wait: true
}

try {    
    parallel map   
}catch(Exception e) {
    e.getCauses().each {
        echo "${it.getShortDescription()}"
     }
}

Actual map creation code is dynamic and looks something like this -

map["${substage}_${subdir}"] = {build job: "ci_pipeline_${substage}${jobSuffix}", parameters: downstreamParams[subdir], propagate: true}

When I execute this, I expect that stages 2 and 3 will fail (simulated) and when I catch the Exception, I get a org.jenkinsci.plugins.workflow.steps.FlowInterruptedException with just one cause element in the cause list -

[Pipeline] Start of Pipeline
[Pipeline] parallel
[Pipeline] { (Branch: second)
[Pipeline] { (Branch: third)
[Pipeline] build (Building childPipeline1)
Scheduling project: childPipeline1
[Pipeline] build (Building childPipeline2)
Scheduling project: childPipeline2
Starting building: childPipeline1 #1333
Starting building: childPipeline2 #148
[Pipeline] }
Failed in branch third
[Pipeline] }
Failed in branch second
[Pipeline] // parallel
[Pipeline] echo
childPipeline2 #148 completed with status FAILURE (propagate: false to ignore)
[Pipeline] End of Pipeline
Finished: SUCCESS

However, when I throw the exception i.e., throw e, I can see all the failure parallel jobs in the Jenkins web console log as below -

[Pipeline] Start of Pipeline
[Pipeline] parallel
[Pipeline] { (Branch: second)
[Pipeline] { (Branch: third)
[Pipeline] build (Building childPipeline1)
Scheduling project: childPipeline1
[Pipeline] build (Building childPipeline2)
Scheduling project: childPipeline2
Starting building: childPipeline1 #1296
Starting building: childPipeline2 #126
[Pipeline] }
Failed in branch third
[Pipeline] }
Failed in branch second
[Pipeline] // parallel
[Pipeline] End of Pipeline
childPipeline2 #126 completed with status FAILURE (propagate: false to ignore)
childPipeline1 #1296 completed with status FAILURE (propagate: false to ignore)
Finished: FAILURE

I am trying to get the result -

childPipeline2 #126 completed with status FAILURE (propagate: false to ignore)

childPipeline1 #1296 completed with status FAILURE (propagate: false to ignore)

and extract just the failed pipeline(s) name and the respective build numbers e.g., childPipeline1: 126 and childPipeline2: 1296 as in the Jenkins web console log (i.e., with all causes; 2 in this example) into a variable in my scripted pipeline.

Please help/guide me with this.

TIA.

Dynamic parallel map runner impl (based on @Yuri G's answer) is as follows -

map["${substage}_${subdir}"] = { try { build job: "ci_pipeline_${substage}${jobSuffix}", parameters: downstreamParams[subdir], propagate: true }catch(Exception e) { e.getCauses().each {echo "${it.getShortDescription()}"}}

Solution

You can catch the error in every branch of parallel execution:

import hudson.model.*
import groovy.json.*

def map = [:]
map['second'] = {
  try {
    build job: "childPipeline1", propagate: true, wait: true    
  }catch(Exception e) {
    e.getCauses().each {
      echo "${it.getShortDescription()}"
    }
    throw e
  }

}

map['third'] = {
  try{
    build job: "childPipeline2", propagate: true, wait: true
  }catch(Exception e) {
     e.getCauses().each {
       echo "${it.getShortDescription()}"
     }
  throw e
  }
}

try {    
  parallel map   
}catch(Exception e) {
  e.getCauses().each {
    echo "${it.getShortDescription()}"
  }
}

Answered By - Yuri G.