如何在Jenkins声明式管道中多Agent循环执行同一任务？

Nice catch on the redundant code—let's refactor this to be cleaner and easier to maintain! The core idea is to define your target node labels in a list, then loop through that list to dynamically generate each stage, eliminating the need to repeat the same stage structure over and over.

Basic Sequential Loop Version #

This will run each clone stage one after another:

pipeline {
    agent none
    stages {
        script {
            // Define all your target node labels in one place
            def targetNodes = ['win10x64-b-ut', 'win81x64-b-ut']
            
            targetNodes.each { nodeLabel ->
                stage("Clone on ${nodeLabel}") {
                    agent { node { label nodeLabel } }
                    steps {
                        gitClone()
                    }
                }
            }
        }
    }
}

Key Details: #

• We centralize all node labels in the targetNodes list—adding a new node later is as simple as inserting a new string here, no more copying entire stage blocks.
• The script block is required in declarative pipelines to use Groovy's dynamic looping features (since declarative is static by default).
• Each iteration generates a stage with a unique name (using the node label) and targets the correct agent automatically.

Optional Parallel Execution Version #

If your Jenkins environment allows running these jobs simultaneously, you can refactor to run the clone stages in parallel to save time:

pipeline {
    agent none
    stages {
        stage('Parallel Clone on All Nodes') {
            parallel {
                script {
                    def targetNodes = ['win10x64-b-ut', 'win81x64-b-ut']
                    targetNodes.each { nodeLabel ->
                        // Dynamically create a parallel branch for each node
                        "${nodeLabel}" {
                            agent { node { label nodeLabel } }
                            steps {
                                gitClone()
                            }
                        }
                    }
                }
            }
        }
    }
}

This wraps all the clone operations into a single parent stage and runs them in parallel, which is great if you have the resources to speed up your pipeline.

内容的提问来源于stack exchange，提问作者vmwgeek