LinuxDay 2016 GitLab Workshop

Da PNLUG.
Versione del 26 ott 2016 alle 12:39 di Amonthedeamon (Discussione | contributi)

(diff) ← Versione meno recente | Versione attuale (diff) | Versione più recente → (diff)

First of all you need to install docker, we'll use Ubuntu 16.04 as base system, and you can just follow some simple instruction provided by digital ocean

For sake of completeness we provide the instruction here:

sudo apt-get update
sudo apt-get upgrade
sudo apt-key adv --keyserver hkp://p80.pool.sks-keyservers.net:80 --recv-keys 58118E89F3A912897C070ADBF76221572C52609D
echo "deb https://apt.dockerproject.org/repo ubuntu-xenial main" | sudo tee /etc/apt/sources.list.d/docker.list
sudo apt-get update
sudo apt-get install -y docker-engine
sudo systemctl status docker
sudo usermod -aG docker $(whoami)

Now logout and login back again to apply the new group policy to your user.

The easier way to that everything went fine, is to run the sample hello-world container:

docker run hello-world

Now that docker is ready to go, we can start with GitLab. The official instruction can be found here

In one single command line:

docker run --detach \
    --hostname gitlab \
    --publish 443:443 --publish 80:80 --publish 22:22 \
    --name gitlab \
    --restart always \
    --volume /srv/gitlab/config:/etc/gitlab \
    --volume /srv/gitlab/logs:/var/log/gitlab \
    --volume /srv/gitlab/data:/var/opt/gitlab \
    gitlab/gitlab-ce:latest

Please note that we choose just gitlab as hostname. This will ease container connection later.

Play around with GitLab by opening your browser to http://localhost : the first time you connect the system will ask you to configure a password for the root user.


Now we want to setup GitLab CI which is a bit harder

The official docs can be found here:

First of all we need to create a runner. A GitLab runner is the worker which will be called by GitLab CI to execute the CI commands.

docker run -d --name gitlab-runner --restart always \
  -v /var/run/docker.sock:/var/run/docker.sock \
  -v /srv/gitlab-runner/config:/etc/gitlab-runner \
  --link gitlab \
  gitlab/gitlab-runner:latest

Please note that we connect this with the previously create GitLab instance with --link parameter, which, in simple words, allow the gitlab-runner container to resolve gitlab hostname in the IP address of the container.

We also provision the docker socket because later docker will be used to execute the CI script.

Now is time to connect the two: in GitLab go into the admin area, and click on Overview -> Runners. Copy the registration token into the clipboard, you'll use it in a minute!

At shell prompt, run the follwing command, which will execute the gitlab-runner registration wizard:

docker exec -it gitlab-runner gitlab-runner register

Enter the follwing parameters:

  • http://localhost
  • paste the token
  • description: docker-runner
  • executor: docker
  • default docker image ruby:2.1

The executor is the way gitlab-runner will execute the GitLab CI script. Choosing docker will execute the CI script inside a container (which will be configured by the CI script itself). A simpler setup is the shell executor which will be execute the CI script locally in the docker-runner container.

However, due the fact that we don't have a real DNS setup, we need to configure gitlab runner to link the container it creates as executor to the GitLab container.

Connect to the runner container and open a shell in it:

docker exec -it gitlab-runner bash

Open with your favorit editor /etc/gitlab-runner/config.toml

In the [runners.docker] section:

	links = [ "gitlab" ]

Easy to guess, this will just add --link gitlab to the docker command line of the executor.

Everything is now ready for you first CI project!

Create a project, e.g. helloworld, with the following helloworld.sh

#!/bin/bash
echo hello GitLab CI world
exit 0

Now create a .gitlab-ci.yml:

image: busybox

job:
    script:
        - chmod +x helloworld.sh
        - ./helloworld.sh

Here it is:

  • this will instruct the runner to use the busybox docker image as executor
  • once the container starts, it will clone our project (this is where the --link will come into)
  • and it finally run the script for the CI

Make it fail!

The above CI will be successfull because all the commands return a zero exit code. Is preatty easy to break the CI: just edit the code, directly from GitLab web interface if you want, to exit 1, commit and.. you break the build!

Fixin it is left as an excercise to the reader.




Code Review Steps (s: senior, j: junior):

  • s: helloworld, con exit 0
  • s: .gitlab-ci.yml
  • s: creare issue con modifica
  • assegnare a junior
  • j: new branch
  • j: modifica readme + [ci skip]
  • j: modifica exit 1 + this closes #1
  • j: CI break build
  • j: submit merge request + assign to s:
  • s: code review
  • j: fix issue with new branch
  • auto-close issue with merge