DemoPipeline Durability

What Is Pipeline Durability?

Pipeline durability in Jenkins controls how pipeline state is written to disk so that builds can recover after a crash or restart. Choosing the right durability level helps you balance:

• Resilience: Ability to resume an in-progress build.
• Performance: Disk I/O overhead when recording pipeline checkpoints.

Durability Levels

Configuring Durability Settings

You can override the default durability level at three scopes:

1. Global (all pipelines)
2. Branch (multibranch projects)
3. Per-Pipeline (individual job)

1. Global Configuration

To set the default durability level for every pipeline:

1. Navigate to Manage Jenkins → Configure System.
2. Search for “Pipeline Speed/Durability.”
3. Choose your preferred default level.

In the same screen, use the dropdown to finalize your choice:

2. Branch-Level (Multibranch Projects)

For multibranch pipelines, you can assign durability per branch:

1. Go to your project’s Configure page.
2. Under Repository Sources → Property strategy, choose Named branches.
3. Add a property for main (or master) with MAX_SURVIVABILITY.
4. Set PERFORMANCE_OPTIMIZED (or another level) for all other branches.

3. Per-Pipeline Job

To set durability for a single pipeline job:

1. Open the job and click Configure.
2. Scroll to Pipeline Speed and Durability.
3. Select the desired durability level from the dropdown.

Live Demo

Demo 1: MAX_SURVIVABILITY

This demo writes numbers to a file once per second. With MAX_SURVIVABILITY, the build will resume after a Jenkins restart.

pipeline {
    agent any
    stages {
        stage('Durability Test Loop') {
            steps {
                script {
                    sh 'touch numbers.txt'
                    for (int i = 0; i < 100; i++) {
                        sh "echo ${i} >> numbers.txt"
                        sleep 1
                    }
                }
            }
        }
    }
}

1. Save and start the build.
2. On the controller host, kill Jenkins:

ps aux | grep -i jenkins.war
kill -9 <JENKINS_PID>
systemctl status jenkins

3. After Jenkins restarts, the console shows:

[Pipeline] sleep
Sleeping for 1 sec
[Pipeline] sh
+ echo 39
[Pipeline] sleep
Sleeping for 1 sec
Resuming build at Sun Nov 10 17:22:47 UTC 2024 after Jenkins restart
No need to sleep any longer
Ready to run at Sun Nov 10 17:22:48 UTC 2024
[Pipeline] sh
+ echo 40

Build execution continues from where it left off.

Demo 2: PERFORMANCE_OPTIMIZED

Switch the same job to PERFORMANCE_OPTIMIZED, rerun, and repeat the kill/restart:

ps aux | grep -i jenkins.war
kill -9 <JENKINS_PID>
systemctl status jenkins

After restart, you’ll see:

ERROR: Cannot resume build because FlowNode 42 for FlowHead 1 could not be loaded. This is expected when using the PERFORMANCE_OPTIMIZED durability setting and Jenkins is not shut down cleanly. Consider switching to the MAX_SURVIVABILITY setting to prevent this.
Finished: FAILURE

Conclusion

• MAX_SURVIVABILITY: Best for mission-critical pipelines that must survive crashes.
• PERFORMANCE_OPTIMIZED: Ideal when speed is paramount and resumability is not required.
• SURVIVABLE_NON_ATOMIC: Use as a balanced choice between safety and performance.

Links and References

• Jenkins Pipeline Documentation
• Multibranch Pipeline
• Managing Jenkins Configuration