Toolchain

Basic Continuous Integration (CI) Scenario

The pipeline examples make use of Compuware's and other plugins to implement the following process and scenario. They can be used to create a standardized approach for implementing a CI process for mainframe development. The scenario(s) and code are based on requests and requirements from customers who already have started implementing their own pipeline and asked Compuware for advice. Regardless of the tool in use to implement a CI strategy, the process we encounter or have helped implement at our customers always looks similar to what we describe here.

The code we share here to implement these process steps, shows Compuware's example approach to these requirements. It does not necessarily reflects actual customer implementations, and is - in parts considerably - different than those implementations.

The Jenkins examples will help mainframe developers to familiarize themselves with Groovy and its concepts, as well as demonstrate some of the idiosyncrasies of the Jenkins Groovy dialect that we come across.

The Azure DevOps Pipeline example, and underlying Powersheel scripts, illustrate that you are not limited to Jenkins when implementing a mainframe-based CI process.

Developer's tasks

Step 1 - Checking out Code

In Topaz a developer checks out a set of sources, copybooks and other components required to fulfill a specific requirement.

Step 2 - Creating/Maintaining Tests

In Topaz Topaz for Total Test the developer creates or modifies a set of (virtualized or non virtualized) tests for the modified programs. In order to share the tests between development teams and to use them in Jenkins, the Topaz for Total Test projects are stored and administered using Git.

Note

Any Git server (e.g. GitHub, Bitbucket, GitLab) may be used as centralized repository server for Topaz for Total Test assets.

Step 3 - Promote Code

In Topaz, once the developer has finished working on the code, they promote their changes from the DEV level to QA, the next level in the application's life cycle.

Note

The IPSW life cycle of the application used for these examples consists of three parallel development paths (DEV1, DEV2, and DEV3). At certain points in the implementation you will need to know the actual path of the life cycle that the developer is working in.

ISPW Life Cylce

Step 4 - Trigger an ISPW webhook

In ISPW, the promotion from the DEV level to the QA level in ISPW automatically creates an ISPW set containing all components that are part of this specific promotion. Once the promotion completes, ISPW triggers a Jenkins job that executes a series of automated steps. The Jenkins job is triggered via a Webhook. The webhook also passes Parameters to the Jenkins job to control the execution.

Step 5 - Trigger a CI pipeline job

In Jenkins or similar a pipeline job or build gets triggered automatically.

CI pipeline job

While there might be differences in details when it comes to the individual implementation examples, certain "stages" are common, as depicted at the top of this page. We provide some examples here:

Step 1 - Retrieve mainframe sources

The ultimate goal of this process is to use SonarQube quality gates to determine if the promoted code is ready to be released. For SonarQube to work, it needs the source code. The source code will be downloaded in one of the early steps (often the first step) of the pipeline.

Step 2 - Retrieve Topaz for Total Test Tests

Topaz for Total Test scenarios and assets are stored in a Git server repository and need to be downloaded to the Jenkins workspace.

Step 3 - Execute Tests

This step will execute the downloaded these tests. Determining which tests to download or which tests to execute may vary based on the requirements of the individual implementation. The examples show how to download every Topaz for Total Test project and execute only those test scenarios that belong to the programs that have been promoted by the developer.

Step 4 - Retrieve Code Coverage statistics from the mainframe

The tests will collect code coverage data and store the results in a Xpediter Code Coverage repository dataset on the mainframe. To be able to pass the results to SonarQube, they need to be extracted from the mainframe repository and downloaded to the mainframe.

Step 5 - Pass results to SonarQube

The sources, test execution results, and code coverage data will be passed to SonarQube for analysis.

Step 6 - Check Quality Gate results

The analysis will be checked against a SonarQube quality gate definition. Based on the result of the quality gate check (pass or fail) further actions will be taken.

Step 6a - On Failure - Regress code in ISPW

If the quality gates fail, usually the code that has been promoted will be regressed (and the responsible personnel will be informed about the failure)-

Step 6b - On success - Trigger a CD release

If the quality gate pass the code is ready to be passed onward in the application life cycle. The Jenkins example will trigger a release in XLRelease, while the Azure DevOps example will trigger an Azure DevOps release pipeline.