COBOL + GitHub + CloudBees Jenkins = Mainframe CI/CD
Overview: Compuware DevOps Specialist Stuart Ashby will present a workshop at the CloudBees Connect virtual event, showing how COBOL source code can be branched, merged and deployed using GitHub, CloudBees and Compuware products.
Why would a mainframe-centric company be attending a CloudBees event?
Automation has always been prevalent on the mainframe platform, using tools that have been designed and developed for the z/OS operating system. Since that advent of the DevOps movement, the lines have blurred significantly. Jenkins as a product name appeared in 2011 and since then has become a proprietary eponym (like Hoover for vacuuming).
Compuware recognised that the Jenkins tool was a market leader for automation and saw an opportunity to establish an open borders pipeline approach to make the mainframe just another server in the datacentre.
Now in 2020, Compuware is participating in a relevant way with CloudBees and other vendors to demonstrate that source code can be stored in your enterprise Git repository and built and deployed on the mainframe by a platform native tool (no emulation, no tool porting), all orchestrated by the market’s leading Continuous Integration tool.
So, what would I consider as relevant content? In my role at Compuware, I travel (pre COVID-19) around our customer base to discuss mainframe DevOps. One of the most common approaches to including the mainframe in a DevOps strategy is to re-use the approach that has worked successfully in the non-mainframe area of an enterprise. This has been demonstrated with SonarQube providing the analysis of COBOL source code.
The most recent initiative I see is that organisations want the following recipe:
• Take a single version control system for all source code
• Add the build of the source on the native platform (no emulation)
• Push and commit to version control to trigger a multibranch pipeline (standard Jenkins functionality)
• Execute SonarQube quality checks during the pipeline, followed by the appropriate unit tests, and any other automation required
• Finally, invoke a peer review, with a pull request, so that the multibranch pipeline can carry out the next level of automation (a deployment and more testing)
That recipe is the demo I plan to show at CloudBees Connect, using familiar user interfaces (such as the Eclipse-based IDE of Topaz Workbench) and familiar tools for version and quality control of COBOL source code (which ultimately becomes a COBOL application on a mainframe). In my opinion, this is the simplicity that developers need to stay in the flow of developing whilst Jenkins and other automation do the heavy lifting.
I invite you to see for yourself in my CloudBees Connect virtual workshop session, “COBOL + GitHub + CloudBees Jenkins = Mainframe CI/CD,” May 19 at 3 PM EDT in North America, and May 20 at 3 PM CET in Europe.