尝试使用 Jenkins 管道脚本克隆 github 存储库时抛出“git：找不到命令”

Question

New to CI/CD building jenkins pipeline.

Use case: copy files from a git repo to AWS S3 bucket. We run Jenkins agents in Kube.netes pods.

So far so good. I am able to connect to Hashicorp vault in the Jenkins pipeline and able to authenticate to our aws account successfully. But when I try to use "git" command to clone the desired repo, it doesn't like git command.

Not sure how to install git on the jenkins agent as part of the jenkins pipeline

this is the successfully running code except the "git command" How to download git in this jenkins pipeline before I clone a repository using 'git" commands

   @Library('enterprise-pipeline-library@master') _
  //Set AWS and Jenkins build properties/parameters
  def agent = "service-deployment-${UUID.randomUUID().toString()}"

  //Container templates

  podTemplates(

      label: agent,

      containers: [
            containerTemplate(name: 'aws-cli', image: 'artifactory.xxx/amazon/aws-cli:2.2.32', 
       ttyEnabled: true, command: 'cat'),
         containerTemplate(name: 'vault', image: 'artifactory.xxx/ease/hashicorp-vault:latest', ttyEnabled: true, command: 'cat')
],              volumes: [hostPathVolume(hostPath: '/var/run/docker.sock', 
     mountPath: 
        '/var/run/docker.sock')],
        imagePullSecrets: ['xxx']
   ){ 
        node(agent) {
                stage("Checkout Code") {
                 echo "hello"
        }
            withCredentials([
                string(credentialsId: 'VAULT_ROLE_ID_UAT', variable: 'VAULT_ROLE_ID_UAT'),
                string(credentialsId: 'VAULT_SECRET_ID_UAT', variable: 'VAULT_SECRET_ID_UAT'),
                string(credentialsId: 'VAULT_ROLE_ID_PCI_UAT', variable: 
              'VAULT_ROLE_ID_PCI_UAT'),
                string(credentialsId: 'VAULT_SECRET_ID_PCI_UAT', variable: 
          'VAULT_SECRET_ID_PCI_UAT'
            ]) {

            stage("Vault Creds for source environment") {                   

                    container('vault') {

                        sh '''

                        export VAULT_TLS_SERVER_NAME=hcvault.xxx
                        export VAULT_ADDR=https://xxxx
                        export VAULT_NAMESPACE=cloud 
                        export ROLE_ID=${VAULT_ROLE_ID_PCI_UAT}
                        export SECRET_ID=${VAULT_SECRET_ID_PCI_UAT}
                        AWS_Account_ID=xxx
                        VAULT_CRED=vault.cred
                        vault write auth/approle/login role_id=$ROLE_ID secret_id=$SECRET_ID  > $VAULT_CRED 2> /dev/null
                        export VAULT_TOKEN=$( awk '/token / {print $2}' $VAULT_CRED ) 
                        AWS_CRED=aws.cred
                        vault read aws/creds/${AWS_Account_ID}-VaultAssumeRole  > $AWS_CRED 2> /dev/null 
                        echo "Removing Vault Credentials..."
                        rm -f $VAULT_CRED 

                        '''
                    } //container                   

             } //stage vault creds

            stage('Sending the files to S3 bucket') {
                container('aws-cli') {                

                        sh '''                          

                        JOB_TITLE=`echo ${JOB_BASE_NAME} | sed 's/ /_/g'`                          

                        set +x

                        AWS_CRED=aws.cred
                        export AWS_ACCESS_KEY_ID=$( awk '/access_key / {print $2}' $AWS_CRED )
                        export AWS_SECRET_ACCESS_KEY=$( awk '/secret_key / {print $2}' $AWS_CRED )
                        export AWS_SESSION_TOKEN=$( awk '/security_token / {print $2}' $AWS_CRED )
                        env | grep AWS

                        set -x

                        git branch: 'main', credentialsId: 'system-id', url: 'https://gitlab.xxx/AMAZON_myaccount/myproject.git'

                        if [ "$sourceEnv" == "UAT_PCI" ]
                        then

                            echo "copying the file to the UAT USE1 bucket"                           

                            #aws s3 cp 7day_rate/pfm_data.json s3://uatpci-pfm-data/data/test2/                              

                            echo "copying the file to the UAT USW2 bucket"
                            #aws s3 cp 7day_rate/pfm_data.json s3://uatpci-pfm-data-usw2/data/test/                          

                        fi 

                        '''

                    }
                }
        } // withcredentials
} // node agent

}

This is the error complaining about "git" command not found. I understand that git has to be downloaded before we use the git branch command to clone the repo. But how can I do a silent install of git on the jenkins agent?

Answer 1

You can follow the git installation instructions from here https://github.com/git-guides/install-git#debianubuntu

You will also need to change the git command. Your code has:

git branch: 'main', credentialsId: 'system-id', url: 'https://gitlab.xxx/AMAZON_myaccount/myproject.git'

This is jenkins pipeline syntax, so it shouldn't be inside a shell script.

I would move the git shell before the shell step, because each shell step opens a new terminal session so the variables you set up inside a shell step only exist during the invocation of that step. If you broke your code into a shell step, a git step and then a shell step, you would have to setup your env variables twice.