[JENKINS-56844] timeout error not propagated when sh step with returnStatus not finished

When a long running sh step with returnStatus is running, timeout error is not propagated.

Steps to reproduce:

1. Having this pipeline, I expect echo 'Continue ...' to never be executed because the internal sh step takes more than the timeout.

```
node {
stage('stage 1') {
test()
echo 'Continue working after timeout'
}
}

def test() {
timeout(time: 10, unit: 'SECONDS') {
echo 'working ...'
sh script: 'sleep 15'//, returnStatus: true
}
}

```

And that works.

2. If the same pipeline is executed removing the comment, enabling returnStatus: true, the interrupt signal is sent but the Pipeline continues and echo 'Continue working after timeout' is executed finishing with status SUCCESS.

Here is the log:

```
Timeout set to expire in 10 sec
[Pipeline] {
[Pipeline] echo
working ...
[Pipeline] sh
[test-pipeline] Running shell script
+ sleep 15
Sending interrupt signal to process
Cancelling nested steps due to timeout
Terminated
[Pipeline] }
[Pipeline] // timeout
[Pipeline] echo
Continue working after timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // node
[Pipeline] End of Pipeline
Finished: SUCCESS

```

---
Originally reported by pcarballo, imported from: timeout error not propagated when sh step with returnStatus not finished

• status: Open
  
• priority: Major
  
• component(s): pipeline
  
• resolution: Unresolved
  
• votes: 1
  
• watchers: 3
  
• imported: 2025-12-08
  

Raw content of original issue

When a long running sh step with returnStatus is running, timeout error is not propagated.


Steps to reproduce:


1. Having this pipeline, I expect echo 'Continue ...' to never be executed because the internal sh step takes more than the timeout.




node {

    stage('stage 1') {

      test()

      echo 'Continue working after timeout'

    }

}


def test() {

    timeout(time: 10, unit: 'SECONDS') {

        echo 'working ...'

        sh script: 'sleep 15'//, returnStatus: true

    }

}





And that works. 


2. If the same pipeline is executed removing the comment, enabling returnStatus: true, the interrupt signal is sent but the Pipeline continues and echo 'Continue working after timeout' is executed finishing with status SUCCESS.


Here is the log:




Timeout set to expire in 10 sec

[Pipeline] {

[Pipeline] echo

working ...

[Pipeline] sh

[test-pipeline] Running shell script

+ sleep 15

Sending interrupt signal to process

Cancelling nested steps due to timeout

Terminated

[Pipeline] }

[Pipeline] // timeout

[Pipeline] echo

Continue working after timeout

[Pipeline] }

[Pipeline] // stage

[Pipeline] }

[Pipeline] // node

[Pipeline] End of Pipeline

Finished: SUCCESS

Start by reproducing the pipeline with timeout and sh using returnStatus: true, then compare it with the version where returnStatus is disabled. Check the timeout and sh step behavior around the interrupt and nested-step cancellation. Done means the timeout error is propagated and the following echo does not execute.

Written by the indexing model from the issue text.