代码之家 › 专栏 › 技术社区 › thclpr

仅使用列表中的最后一项使用声明性管道执行并行阶段

jenkins-pipeline jenkins

thclpr · 技术社区 · 8 年前

我不确定我在这里做错了什么,因为目前当我尝试在列表上迭代时,阶段的创建似乎很好,但在执行shellscript时,使用的值始终是列表的最后一项:

工作管道:

pipeline {
    agent any
    stages {
        stage('set servers') {
            steps {
                script {
                    my_list = ['server1','server-2','server-3']
                }
            }
        }
        stage('Execute then') {
            parallel {
                stage('shouter') {
                    steps {            
                        script {
                            shouter = [:]
                            script {
                                for(i in my_list) {
                                    shouter["${i}"] = {
                                        echo "standupandshout.sh ${i}"
                                    }
                                }
                            }                            
                            parallel shouter
                        }
                    }
                }
            }
        }
    }
}

输出:

控制台输出:

Replayed #4
Running in Durability level: MAX_SURVIVABILITY
[Pipeline] node
Running on Jenkins in /var/lib/jenkins/workspace/test
[Pipeline] {
[Pipeline] stage
[Pipeline] { (set servers)
[Pipeline] script
[Pipeline] {
[Pipeline] }
[Pipeline] // script
[Pipeline] }
[Pipeline] // stage
[Pipeline] stage
[Pipeline] { (Execute then)
[Pipeline] parallel
[Pipeline] [shouter] { (Branch: shouter)
[Pipeline] [shouter] stage
[Pipeline] [shouter] { (shouter)
[Pipeline] [shouter] script
[Pipeline] [shouter] {
[Pipeline] [shouter] script
[Pipeline] [shouter] {
[Pipeline] [shouter] }
[Pipeline] [shouter] // script
[Pipeline] [shouter] parallel
[Pipeline] [server1] { (Branch: server1)
[Pipeline] [server-2] { (Branch: server-2)
[Pipeline] [server-3] { (Branch: server-3)
[Pipeline] [server1] echo
[server1] standupandshout.sh server-3
[Pipeline] [server1] }
[Pipeline] [server-2] echo
[server-2] standupandshout.sh server-3
[Pipeline] [server-2] }
[Pipeline] [server-3] echo
[server-3] standupandshout.sh server-3
[Pipeline] [server-3] }
[Pipeline] [shouter] // parallel
[Pipeline] [shouter] }
[Pipeline] [shouter] // script
[Pipeline] [shouter] }
[Pipeline] [shouter] // stage
[Pipeline] [shouter] }
[Pipeline] // parallel
[Pipeline] }
[Pipeline] // stage
[Pipeline] }
[Pipeline] // node
[Pipeline] End of Pipeline
Finished: SUCCESS

所需输出:

[Pipeline] [server1] echo
[server1] standupandshout.sh server-1
[Pipeline] [server1] }
[Pipeline] [server-2] echo
[server-2] standupandshout.sh server-2
[Pipeline] [server-2] }
[Pipeline] [server-3] echo
[server-3] standupandshout.sh server-3

1 回复 | 直到 8 年前

thclpr 8 年前

这是由于groovy闭包以及它们所包含的代码得到计算时造成的。 http://blog.freeside.co/2013/03/29/groovy-gotcha-for-loops-and-closure-scope/

当闭包运行时,绑定到变量i的值是它在循环的最终迭代上的值,而不是创建闭包的迭代上的值。闭包的作用域引用了i,在执行任何闭包时,i是5。

循环体的局部变量的行为与此不同,显然是因为每个闭包范围都包含对不同变量的引用

这就是为什么你的舞台名称是可以的,但你的价值不是。

解决方案是什么?我们应该一直使用。每个而不是一个for循环?嗯,我有点喜欢循环,在许多情况下,可能会有内存利用率的差异(不要认为这意味着循环“更好”或“更高效”)。

如果您只需在封闭体一切正常

def fns = []
for (i in (1..5)) {
    def myi = i
    def isq = i * i
    fns << {->
        println "$myi squared is $isq"
    }
}
fns.each { it() }

所以这应该是可行的:

script {
  shouter = [:]
  for(i in my_list) {
    def val = i
    shouter["${i}"] = {
      echo "standupandshout.sh ${val}"
    }
  }
  parallel shouter
}