In this tutorial we will setup a Basic CI (Continuous Integration) Pipeline on Gitlab.
The intention of this post is just to touch on the basics, and in my next post, we will setup a ci/cd pipeline to deploy a python flask api with postgres on heroku.
The code for this example is available on my gitlab.com repo.
- Setup a Basic CI Pipeline on Gitlab (This Post)
- Setup a Gitlab CI/CD Pipeline to Deploy your Python API to Heroku
- Setup a Gitlab CI Runner on your own Server
The aim for this is every time there is a commit made to the master branch, that the jobs defined by the
.gitlab-ci.yml will be executed and will only pass if exit code 0 has been returned on the scripts.
The jobs gets executed on gitlab runners which is hosted with Giltab. Important to note is that every job runs independently from each other.
In this pipeline we will have 2 basic jobs, each job execute a set of scripts:
$ echo "this is building" $ hostname $ mkdir builds $ touch builds/data.txt $ echo "true" > builds/data.txt
$ echo "this is testing" $ hostname $ test -f builds/data.txt $ grep "true" builds/data.txt
From a newly created repository which i've cloned to my workspace, create the config:
$ touch .gitlab-ci.yml
The config for above yaml file:
stages: - build - test build: stage: build script: - echo "this is building" - hostname - mkdir builds - touch builds/data.txt - echo "false" > builds/data.txt artifacts: paths: - builds/ test: stage: test script: - echo "this is testing" - hostname - test -f builds/data.txt - grep "true" builds/data.txt
- We define 2 stages for this pipeline: build and test
- We provide context of each job, the stage, the script (commands that will be executed in the lifecycle of the runner) and artifacts (artifacts will be the content that will be transferred, as each job runs in a different runner/container)
Note that I deliberately made a mistake so that my pipeline can fail. I populated the content "false" into the
builds/data.txt file from the build job and grep for the word "true" on the test job, so this job will fail.
Save the content to the config file, add, commit and push to master:
$ git add .gitlab-ci.yml $ git commit -m "add gitlab-ci config" $ git push origin master
From the Gitlab UI, if you head over to CI/CD -> Pipelines, you should see your pipeline running:
When you select the Pipeline ID, you should be presented with the jobs available in your pipeline:
Select Jobs, and you should see an overview of your jobs. At this moment we can see that the build job has completed, and that the test job is busy running:
Shortly thereafter the status of the test job should change to failed, select the Job ID and you should see the output:
From the above output it gives you a link to create a new issue, which is quite handy.
Let's go ahead and change the content in the
.gitlab-ci.yml config and push to master:
$ vim .gitlab-ci.yml
Change line 12 from
- echo "false" > builds/data.txt to
- echo "true" > builds/data.txt, the full content of the file:
stages: - build - test build: stage: build script: - echo "this is building" - hostname - mkdir builds - touch builds/data.txt - echo "true" > builds/data.txt artifacts: paths: - builds/ test: stage: test script: - echo "this is testing" - hostname - test -f builds/data.txt - grep "true" builds/data.txt
Commit and push to master:
$ git add .gitlab-ci.yml $ git commit -m "change content in script" $ git push origin master
When you head over to Pipelines, you will see that the pipeline is busy running, and on the right the commit that we just made:
Select the Pipeline ID, then select Jobs, you should see both jobs succeeded:
Select the Job ID of the test job, and from the output you will see that the job succeeded:
From this output you can also confirm from both jobs, that each job ran in a different runner as the hostnames that was returned to stdout was different.
This was a really basic example to demonstrate Gitlab CI. Some relevant resources to this post: