jenkinsci / jenkinsci/workflow-durable-task-step-plugin
[JENKINS-73734] Timeout of stage does not cause build to fail when running PowerShell with bat
- Dominant language
- Java
- Stars
- 46
- Forks
- 110
- PR merge metrics
- No merged PRs in 30d
Description
Steps to reproduce
Run the following declarative pipeline multiple times:
pipeline {
agent { label 'windows' }
stages {
stage('Stage') {
options {
timeout(time: 5, unit: 'SECONDS')
}
steps {
script {
writeFile(file: 'sleep.bat', text: 'ping -n 100 localhost')
writeFile(file: 'sleep.ps1', text: 'Start-Sleep 100')
// FAILS 50 %
bat 'powershell -File sleep.ps1'
// OK
//bat 'ping -n 100 localhost'
// OK
//bat 'sleep.bat'
// OK
//powershell 'Start-Sleep 100'
// OK
//powershell './sleep.ps1'
}
}
}
}
post {
always {
echo 'always'
}
success {
echo 'success'
}
aborted {
echo 'aborted'
}
failure {
echo 'failure'
}
unsuccessful {
echo 'unsuccessful'
}
}
}
Expected behavior
The timeout is honored.
The build fails on every run.
The failed builds are displayed "gray".
In the post actions, "aborted" and "unsuccessful" is echoed.
Actual behavior
The timeout is honored.
The build fails only around 50% of the runs. On about every other run, the build is marked green although it ran into timeout. In these cases, the post actions echo "successful".

Log of correct gray build:
```
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Stage)
[Pipeline] timeout
Timeout set to expire in 5 sec
[Pipeline] {
[Pipeline] script
[Pipeline] {
[Pipeline] writeFile
[Pipeline] writeFile
[Pipeline] batC:\Jenkins\workspace\TimeoutDebug>powershell -File sleep.ps1
Cancelling nested steps due to timeout
Sending interrupt signal to process
script returned exit code -1
[Pipeline] }
[Pipeline] // script
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Declarative: Post Actions)
[Pipeline] echo
always
[Pipeline] echo
aborted
[Pipeline] echo
unsuccessful
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // node
[Pipeline] End of Pipeline
Timeout has been exceeded
org.jenkinsci.plugins.workflow.actions.ErrorAction$ErrorId: 11ad6797-e9a8-4ec5-80f7-5ac4925f7fd0
Finished: ABORTED
```
Log of incorrect green build:
```
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Stage)
[Pipeline] timeout
Timeout set to expire in 5 sec
[Pipeline] {
[Pipeline] script
[Pipeline] {
[Pipeline] writeFile
[Pipeline] writeFile
[Pipeline] batC:\Jenkins\workspace\TimeoutDebug>powershell -File sleep.ps1
Cancelling nested steps due to timeout
Sending interrupt signal to process
Batchvorgang abbrechen (J/N)?
[Pipeline] }
[Pipeline] // script
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Declarative: Post Actions)
[Pipeline] echo
always
[Pipeline] echo
success
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // node
[Pipeline] End of Pipeline
Finished: SUCCESS
```
Additional information
I am aware of the "powershell" step and that this is the preferred way of running PowerShell. As you can see in the example pipeline, the bug is not present when using the "powershell" step. Nevertheless I wanted to inform you about this bug, as maybe other executables are affected as well.
---
Originally reported by tracetronic, imported from: Timeout of stage does not cause build to fail when running PowerShell with bat
Raw content of original issue
Steps to reproduce
Run the following declarative pipeline multiple times:
pipeline {
agent { label 'windows' }
stages {
stage('Stage') {
options {
timeout(time: 5, unit: 'SECONDS')
}
steps {
script {
writeFile(file: 'sleep.bat', text: 'ping -n 100 localhost')
writeFile(file: 'sleep.ps1', text: 'Start-Sleep 100')
// FAILS 50 %
bat 'powershell -File sleep.ps1'// OK
//bat 'ping -n 100 localhost'
// OK
//bat 'sleep.bat'
// OK
//powershell 'Start-Sleep 100'
// OK
//powershell './sleep.ps1'
}
}
}
}
post {
always {
echo 'always'
}
success {
echo 'success'
}
aborted {
echo 'aborted'
}
failure {
echo 'failure'
}
unsuccessful {
echo 'unsuccessful'
}
}
}Expected behavior
The timeout is honored.
The build fails on every run.
The failed builds are displayed "gray".
In the post actions, "aborted" and "unsuccessful" is echoed.Actual behavior
The timeout is honored.
The build fails only around 50% of the runs. On about every other run, the build is marked green although it ran into timeout. In these cases, the post actions echo "successful".
Log of correct gray build:
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Stage)
[Pipeline] timeout
Timeout set to expire in 5 sec
[Pipeline] {
[Pipeline] script
[Pipeline] {
[Pipeline] writeFile
[Pipeline] writeFile
[Pipeline] batC:\Jenkins\workspace\TimeoutDebug>powershell -File sleep.ps1
Cancelling nested steps due to timeout
Sending interrupt signal to process
script returned exit code -1
[Pipeline] }
[Pipeline] // script
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Declarative: Post Actions)
[Pipeline] echo
always
[Pipeline] echo
aborted
[Pipeline] echo
unsuccessful
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // node
[Pipeline] End of Pipeline
Timeout has been exceeded
org.jenkinsci.plugins.workflow.actions.ErrorAction$ErrorId: 11ad6797-e9a8-4ec5-80f7-5ac4925f7fd0
Finished: ABORTEDLog of incorrect green build:
[Pipeline] {
[Pipeline] stage
[Pipeline] { (Stage)
[Pipeline] timeout
Timeout set to expire in 5 sec
[Pipeline] {
[Pipeline] script
[Pipeline] {
[Pipeline] writeFile
[Pipeline] writeFile
[Pipeline] batC:\Jenkins\workspace\TimeoutDebug>powershell -File sleep.ps1
Cancelling nested steps due to timeout
Sending interrupt signal to process
Batchvorgang abbrechen (J/N)?
[Pipeline] }
[Pipeline] // script
[Pipeline] }
[Pipeline] // timeout
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Declarative: Post Actions)
[Pipeline] echo
always
[Pipeline] echo
success
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // node
[Pipeline] End of Pipeline
Finished: SUCCESSAdditional information
I am aware of the "powershell" step and that this is the preferred way of running PowerShell. As you can see in the example pipeline, the bug is not present when using the "powershell" step. Nevertheless I wanted to inform you about this bug, as maybe other executables are affected as well.
environment
```
Jenkins 2.462.1
Pipeline: Nodes and Processes Version 1371.vb_7cec8f3b_95e
Microsoft Windows [Version 10.0.19045.4780]
$PSVersionTable
Name Value
---- -----
PSVersion 5.1.19041.4780
PSEdition Desktop
PSCompatibleVersions {1.0, 2.0, 3.0, 4.0...}
BuildVersion 10.0.19041.4780
CLRVersion 4.0.30319.42000
WSManStackVersion 3.0
PSRemotingProtocolVersion 2.3
SerializationVersion 1.1.0.1
```
1 attachment
- [2024-09-06 08_14_50-.png](https://issues.jenkins.io/secure/attachment/63239/2024-09-06%2008_14_50-.png)
> 
Contributor guide
Assessment
This issue has not been assessed yet.