Page tree
Skip to end of metadata
Go to start of metadata

Based on a straightforward and repeatable process, the continuous delivery approach helps with building, testing and releasing software faster. On this page, we will show how you can automate your build, deploy, and testing processes using an automation tool together with the E2E command line tools.
To take full advantage of this scenario, you need the E2E Command Line Compiler which is part of E2E Bridge 7.

Prerequisites

The following will guide you through an continuous delivery example. Here, we are using Jenkins, an open source automation tool, and Git, an open source version control tool. To set up a similar scenario, you need

Approach

  • Set-up a job in Jenkins that checks minutely for changes to your git repository and triggers a build process in case of changes.
  • The build process itself is defined in a Jenkinsfile that is part of the Git repository and contains the following steps:
    1. Building the xUML repository file using the E2E Command Line Compiler.
    2. Deploying the compiled repository using the Bridge Command Line Interface.
    3. Running regression tests on the deployed services with the E2E Regression Test Runner.

Setting Up a Job in Jenkins

First, you need to set up a job in Jenkins that will look for changes to your Git repository regularly and trigger the build process in case of changes.

Create a new Jenkins item: a Multibranch Pipline job.

Specify where to get the sources from: it will be Git in this example.
Select a Git repository and provide credentials if necessary.

Build configuration will be by Jenkinsfile. This file will be explained further below.

Specify the triggers that will lead to execution of this job.
Check Periodically and select 1 (once per) minute. Jenkins will check every minute for changes on the repository and then do what is configured in the Jenkinsfile.

Setting Up a Jenkins Command File

In your Git repository, you need a Jenkinsfile - a Apache Groovy script file that contains the build steps to be performed by Jenkins. The file must be named exactly like that, Jenkinsfile (without extension), and must reside in the root directory of the Git repository.

 Click here to expand the complete code of the Jenkinsfile ...
#!groovy

pipeline {
    agent {
        node {
            label 'Windows'
            customWorkspace "workspace/test-xUML-project"
        }
    }

    options {
        buildDiscarder(logRotator(numToKeepStr: '10', artifactNumToKeepStr: '1'))
        disableConcurrentBuilds()
    }
    
    parameters {
        choice(name: 'E2EC', choices: 'D:/jenkins/userContent/e2ec/e2ec-7.0.0-beta1.jar', description: 'Location of the E2E Compiler')
        choice(name: 'REGTEST', choices: 'D:/jenkins/userContent/RegTestRunner/RegTestRunner-nightly.jar', description: 'Location of the E2E Regression Test Runner')
    }

    stages {
        stage('Build') {
            steps {
                dir('Advanced Modeling/E2ELibrary') {
                    bat """
                        java -jar ${E2EC} -uml uml/librarySQLQuery.xml
                        copy repository\\librarySQLQuery\\librarySQLQuery.lrep libs\\
                        java -jar ${E2EC} -uml uml/useLibrarySQLQuery.xml
                    """
                    archiveArtifacts artifacts: 'repository/useLibrarySQLQuery/UseE2ELibraryExample.rep'
                }
                
                dir('Advanced Modeling/PState') {
                    bat """
                        java -jar ${E2EC} -uml uml/pstatePurchaseOrder.xml
                    """
                    archiveArtifacts artifacts: 'repository/pstatePurchaseOrder/PurchaseOrderExample.rep'
                }
            }
        }
        stage('Deploy') {
            steps {
                dir('Advanced Modeling') {
                    bat '''
                        e2ebridge deploy E2ELibrary/repository/useLibrarySQLQuery/UseE2ELibraryExample.rep -h <Bridge host> -u <user> -P <password> -o overwrite
                        e2ebridge deploy PState/repository/pstatePurchaseOrder/PurchaseOrderExample.rep -h <Bridge host> -u <user> -P <password> -o overwrite
                    '''
                }
            }
        }
        stage('Test') {
            steps {
                dir('Advanced Modeling') {
                    bat """
                        java -jar ${REGTEST} -project PState -suite "QA Tests/Tests" -logfile result.xml -host <Bridge host> -port <port> -username <user> -password <password>
                    """
                }
            }
            post {
                always {
                    junit 'Advanced Modeling/result.xml'
                }
            }
        }
    }
}
Code SnippetDescription
agent {
    node {
        label 'Windows'
        customWorkspace "workspace/test-xUML-project"
    }
}

Section agent defines

  • the name of the agent that should execute the build
  • the common workspace directory on this agent for all branches within the Git repository. If not specified, Jenkins will create a separate workspace for every branch.
options {
    buildDiscarder(logRotator(numToKeepStr: '10', artifactNumToKeepStr: '1'))
    disableConcurrentBuilds()
}


Section options defines

  • the count of build logs to keep
  • the count of build artifacts to keep
  • whether concurrent builds are allowed
parameters {
    choice(name: 'E2EC', choices: 'D:/jenkins/userContent/e2ec/e2ec-7.0.0-beta1.jar', description: 'Location of the E2E Compiler')
    choice(name: 'REGTEST', choices: 'D:/jenkins/userContent/RegTestRunner/RegTestRunner-nightly.jar', description: 'Location of the E2E Regression Test Runner')
}


Section parameters defines parameters to use further below in the script. If you provide multiple choices, Jenkins will generate a dropdown list to select from. The first list item serves as the default value. This default will be selected, if the script is triggered automatically.

A build with newly added parameters will always fail for the first time, because Jenkins needs the first run to add the parameters to the pipeline configuration.

stages {
    stage('Build') { 
        [...]
    }
}


In section stages, you can define named build stages. If the processing of one stage fails, the subsequent stages will not be processed.
steps {
    dir('Advanced Modeling/E2ELibrary') {
    bat """
        java -jar ${E2EC} -uml uml/librarySQLQuery.xml
        copy repository\\librarySQLQuery\\librarySQLQuery.lrep libs\\
        java -jar ${E2EC} -uml uml/useLibrarySQLQuery.xml
    """
    archiveArtifacts artifacts: 'repository/useLibrarySQLQuery/UseE2ELibraryExample.rep'
} 


 In section steps, you can define the tasks to process in this stage, e.g.

  • dir: to change the active directory within the Git repository
  • bat: to execute batch commands (Windows), e.g.
    1. Call the library build.
    2. Copy the library repository to the libs folder.
    3. Call the xUML model build.

    Wrap the batch commands in

    • single quotes ('), if Jenkins variables should not be resolved within the batch command
    • double quotes ("), if Jenkins variables should be resolved within the batch command
    • triple single quotes (''') or triple double quotes (""") for multiple line batch scripts
  • archiveArtifacts: to define a list of artifacts (outputs) of this job. These artifacts can be downloaded via the Jenkins console.
java -jar ${E2EC} -uml uml/librarySQLQuery.xml


Call the E2E Command Line Compiler. The location of the compiler is specified via a Jenkins parameter (see above).
copy repository\\librarySQLQuery\\librarySQLQuery.lrep libs\\


Copy the compiled library repository to the libs folder of the project, so it will be used when compiling the usage model. For more details, see Compiling Libraries and Library Usage Models.
e2ebridge deploy E2ELibrary/repository/useLibrarySQLQuery/UseE2ELibraryExample.rep -h <Bridge host> -u <user> -P <password> -o overwrite


Call the E2E Bridge CLI to deploy the compiled service to an E2E Bridge.
dir('Advanced Modeling') {
    bat """
        java -jar ${REGTEST} -project PState -suite "QA Tests/Tests" -logfile result.xml -host <Bridge host> -port <port> -username <user> -password <password>
    """
}
Call the E2E RegTestRunner to perform regression tests on the newly deployed service. The location of the RegTestRunner is specified via a Jenkins parameter (see above).

For more details on Apache Groovy, refer to the Apache Groovy Script Documentation.
You can check the script's syntax before execution using Pipeline Linter, a command line tool that is coming with Jenkins. Refer to the Jenkins Documentation for more information on Linter.

The Build Process

And that's all? Yes, it is. Every time you will push changes to the related Git repository, the multibranch pipeline job will be triggered automatically and perform the defined steps.

On the Jenkins console, you can find a nice overview on each test run:

The Jenkins console log shows the processing in detail:

Some Jenkins Hints

HintDescription
first run of a Jenkins jobThe first run of a Jenkins job may take a little longer, depending on the size of the Git repository that is checked out.
additional stages

You can not only use Jenkins to automate your building, deploying and testing, but also to

  • publish the build result to a delivery endpoint

  • send mails containing the build results to all who are concerned

parameters

You can use parameters with your multibranch pipeline (see also Jenkinsfile above).

  • Parameter definition: Jenkinsfile
  • Parameter value specification: Jenkins job configuration or first item if choices (default value), if triggered automatically
  • Parameter usage: Jenkinsfile
  • The first run of a parameterized pipeline will fail, because Jenkins needs the first run to add the parameters to the pipeline configuration. Same, when parameters change.
concurrent buildsYou can use disableConcurrentBuilds() to prevent the multibranch pipeline to be triggered while another build is still running. Nevertheless, this will not prevent multiple branches within the same job being processed in parallel.
To prevent this, you could e.g. allow certain build steps - that cannot run in parallel - for specific branches only.

Other Useful Tasks to be Automated

TaskDescription
change service preferences

You can use the E2E Bridge CLI to change the preferences of a service, e.g. to set the flag for automatic startup:

e2ebridge preferences <service name> --pref.automaticStartup=true -u <user> -P <password>

You can list all available preferences with

e2ebridge preferences <service name> -u <user> -P <password>
change service settings

You can call the Bridge API with cURL to change the settings of a service:

curl -X PUT --header "Content-Type: application/json" --upload "<JSON settings file>" -G "https://localhost:8080/bridge/rest/services/xuml/<service name>/settings" -u <user>:<password> -k -v 

To execute this PUT, you need a JSON file that contains the complete definition of all service settings. You can obtain this by

  1. executing a GET on the settings of the service
  2. saving the results of this GET to a file
  3. manipulating this file with the setting changes you want to send to the service
  • No labels