In two previous posts, I’ve showed how to configure Jenkins in a container as well as how to add the SQL Change Automation plugin. This post looks at an initial database build in Jenkins. I will build from my local repo, which gives me a bit of a pre-push check on my code.
Note: I already have a SQL Source Control (SOC) project committed to Github that contains the database I want to build. If you haven’t done that, see this post on a basic SOC->GitHub link.
Creating a Jenkins Job
Jenkins calls these build processes jobs, and you see that if you set up Jenkins in a container. By default, there are no jobs, and Jenkins lets you know this.
When I click this, I get a choice of different projects. You can read about these in the Jenkins documentation, but I am just going to use a Freestyle Project. Give it a name and click the type and then OK.
At this point, we need to configure the job. I do like to provide a description, and normally I would set this as a Github project, but in this case, this is really just a local build..
I do want to restrict which agent will run this, since I have a master agent in the Jenkins container. That one can’t see my SQL Server (and doesn’t have tools), so I’ll restrict this to the local agent on my machine.
There are tabs at the top for quick navigation, but this config is really just one long page. You can scroll down as needed. Below the agents, I find the VCS section. I’ll choose git, and then enter the path the agent will need.
When I scroll down to the Build area, I need to add a step. CI is just a step engine to assemble your software, and in this case, I see my Redgate SQL Change Automation plugin.
When I click this, it is added as the first step. There is an implicit “git clone” checkout of code. I need to configure this step with just a few items. First, build in the local folder, rather than some specific area. We’ve already specified an agent location when we configured the Jenkins agent.
Also, you need a package ID, which will be the place where the current state of all objects is stored. Unlike application code, the artifact is really the current version of the objects in source control, not any .exe or .dll.
The only way to build a database project in SQL Server is with an actual SQL Server. In this case, I don’t have any code that would error on LocalDB, so I’ll just use that. I coudl specify my local SQL Server development database if I had the need.
This is a test build, so I also don’t need any SQL Compare options or other switches.
This kicks off a build on my agent. If your agent isn’t connected to the container, you’ll need to do that now. At the bottom left of the page, I see my build working.
The build number is a link, so click that to go to the details of the build. When you get there, you can see the build working by clicking the Console output. As you see, the output is on the right side of the screen. As you can see, my local path worked fine, without any error.
I also got to this part of the process on a plane, with no Internet. While there were some warnings as the plugin and other parts of the process tried to contact the PowerShell Gallery, things worked.
At the end, I see success, which means that the code I wrote in my database, and that was committed to my VCS, is valid. It’s a low bar, but you might be surprised how often someone breaks code by making a strange change to their project.
This is a basic build process, to which I can add other steps. Testing and releasing are further steps, and Jenkins can support you in these.
Now I can move forward with additional development, knowing I have my own build system in my own environment, using Jenkins in a container.