Evaggelos Balaskas - System Engineer

GitLab is my favorite online git hosting provider, and I really love the CI feature (that now most of the online project providers are also starting supporting it).

Archlinux uses git and you can find everything here: Arch Linux git repositories

There are almost 2500 packages there! There are 6500 in core/extra/community (primary repos) and almost 55k Packages in AUR, the Archlinux User Repository.

We are going to use git to retrieve our PKGBUILD from aur archlinux as an example.
The same can be done with one of the core packages by using the above git repo.

So here is a very simple .gitlab-ci.yml file that we can use to build an archlinux package in gitlab

image: archlinux/base:latest

before_script:
    - export PKGNAME=tallow

run-build:
  stage: build
  artifacts:
    paths:
    - "*.pkg.tar.xz"
    expire_in: 1 week
  script:
      # Create "Bob the Builder" !
    - groupadd bob && useradd -m -c "Bob the Builder" -g bob bob
      # Update archlinux and install git
    - pacman -Syy && pacman -Su --noconfirm --needed git base-devel
      # Git Clone package repository
    - git clone https://aur.archlinux.org/$PKGNAME.git
    - chown -R bob:bob $PKGNAME/
      # Read PKGBUILD
    - source $PKGNAME/PKGBUILD
      # Install Dependencies
    - pacman -Syu --noconfirm --needed --asdeps "${makedepends[@]}" "${depends[@]}"
      # Let Bob the Builder, build package
    - su - bob -s /bin/sh -c "cd $(pwd)/$PKGNAME/ && makepkg"
      # Get artifact
    - mv $PKGNAME/*.pkg.tar.xz ./

You can use this link to verify the above example: tallow at gitlab

But let me explain the steps:

• First we create a user, Bob the Builder as in archlinux we can not use root to build a package for security reasons.
• Then we update our container and install git and base-devel group. This group contains all relevant archlinux packages for building a new one.
• After that, we git clone the package repo
• Install any dependencies. This is a neat trick that I’ve found in archlinux forum using source command to create shell variables (arrays).
• Now it is time for Bob to build the package !
• and finally, we move the artifact in our local folder

Ansible is a wonderful software to automatically configure your systems. The default mode of using ansible is Push Model.

That means from your box, and only using ssh + python, you can configure your flee of machines.

Ansible is imperative. You define tasks in your playbooks, roles and they will run in a serial manner on the remote machines. The task will first check if needs to run and otherwise it will skip the action. And although we can use conditional to skip actions, tasks will perform all checks. For that reason ansible seems slow instead of other configuration tools. Ansible runs in serial mode the tasks but in psedo-parallel mode against the remote servers, to increase the speed. But sometimes you need to gather_facts and that would cost in execution time. There are solutions to cache the ansible facts in a redis (in memory key:value db) but even then, you need to find a work-around to speed your deployments.

But there is an another way, the Pull Mode!

Useful Reading Materials

to learn more on the subject, you can start reading these two articles on ansible-pull.

• Jan-Piet Mens

• Stavros Korokithakis

Pull Mode

So here how it looks:

You will first notice, that your ansible repository is moved from you local machine to an online git repository. For me, this is GitLab. As my git repo is private, I have created a Read-Only, time-limit, Deploy Token.

With that scenario, our (ephemeral - or not) VMs will pull their ansible configuration from the git repo and run the tasks locally. I usually build my infrastructure with Terraform by HashiCorp and make advance of cloud-init to initiate their initial configuration.

Cloud-init

The tail of my user-data.yml looks pretty much like this:

...
# Install packages
packages:
  - ansible

# Run ansible-pull
runcmd:
  - ansible-pull -U https://gitlab+deploy-token-XXXXX:YYYYYYYY@gitlab.com/username/myrepo.git 

Playbook

You can either create a playbook named with the hostname of the remote server, eg. node1.yml or use the local.yml as the default playbook name.

Here is an example that will also put ansible-pull into a cron entry. This is very useful because it will check for any changes in the git repo every 15 minutes and run ansible again.

- hosts: localhost

  tasks:
    - name: Ensure ansible-pull is running every 15 minutes
      cron:
        name: "ansible-pull"
        minute: "15"
        job: "ansible-pull -U https://gitlab+deploy-token-XXXXX:YYYYYYYY@gitlab.com/username/myrepo.git &> /dev/null"

    - name: Create a custom local vimrc file
      lineinfile:
        path: /etc/vim/vimrc.local
        line: 'set modeline'
        create: yes

    - name: Remove "cloud-init" package
      apt:
        name: "cloud-init"
        purge: yes
        state: absent

    - name: Remove useless packages from the cache
      apt:
        autoclean: yes

    - name: Remove dependencies that are no longer required
      apt:
        autoremove: yes

# vim: sts=2 sw=2 ts=2 et

Continuous Deployment with GitLab: how to build and deploy a RPM Package with GitLab CI

I would like to automate building custom rpm packages with gitlab using their CI/CD functionality. This article is a documentation of my personal notes on the matter.

[updated: 2018-03-20 gitlab-runner Possible Problems]

Installation

You can find notes on how to install gitlab-community-edition here: Installation methods for GitLab. If you are like me, then you dont run a shell script on you machines unless you are absolutely sure what it does. Assuming you read script.rpm.sh and you are on a CentOS 7 machine, you can follow the notes below and install gitlab-ce manually:

Import gitlab PGP keys

# rpm --import https://packages.gitlab.com/gitlab/gitlab-ce/gpgkey 

# rpm --import https://packages.gitlab.com/gitlab/gitlab-ce/gpgkey/gitlab-gitlab-ce-3D645A26AB9FBD22.pub.gpg

Gitlab repo

# curl -s 'https://packages.gitlab.com/install/repositories/gitlab/gitlab-ce/config_file.repo?os=centos&dist=7&source=script' \
  -o /etc/yum.repos.d/gitlab-ce.repo 

Install Gitlab

# yum -y install gitlab-ce

Configuration File

The gitlab core configuration file is /etc/gitlab/gitlab.rb
Remember that every time you make a change, you need to reconfigure gitlab:

# gitlab-ctl reconfigure

My VM’s IP is: 192.168.122.131. Update the external_url to use the same IP or add a new entry on your hosts file (eg. /etc/hosts).

external_url 'http://gitlab.example.com'

Run: gitlab-ctl reconfigure for updates to take effect.

Firewall

To access the GitLab dashboard from your lan, you have to configure your firewall appropriately.

You can do this in many ways:

• Accept everything on your http service
  # firewall-cmd --permanent --add-service=http

• Accept your lan:
  # firewall-cmd --permanent --add-source=192.168.122.0/24

• Accept only tcp IPv4 traffic from a specific lan
  # firewall-cmd --permanent --direct --add-rule ipv4 filter INPUT 0 -p tcp -s 192.168.0.0/16 -j ACCEPT

or you can complete stop firewalld (but not recommended)

• Stop your firewall
  # systemctl stop firewalld

okay, I think you’ve got the idea.

Reload your firewalld after every change on it’s zones/sources/rules.

# firewall-cmd --reload

success

Browser

Point your browser to your gitlab installation:

http://192.168.122.131/

this is how it looks the first time:

and your first action is to Create a new password by typing a password and hitting the Change your password button.

Login

First Page

New Project

I want to start this journey with a simple-to-build project, so I will try to build libsodium,
a modern, portable, easy to use crypto library.

New project --> Blank project

I will use this libsodium.spec file as the example for the CI/CD.

Docker

The idea is to build out custom rpm package of libsodium for CentOS 6, so we want to use docker containers through the gitlab CI/CD. We want clean & ephemeral images, so we will use containers as the building enviroments for the GitLab CI/CD.

Installing docker is really simple.

Installation

# yum -y install docker 

Run Docker

# systemctl restart docker
# systemctl enable  docker

Download image

Download a fresh CentOS v6 image from Docker Hub:

# docker pull centos:6 

Trying to pull repository docker.io/library/centos ...
6: Pulling from docker.io/library/centos
ca9499a209fd: Pull complete
Digest: sha256:551de58ca434f5da1c7fc770c32c6a2897de33eb7fde7508e9149758e07d3fe3

View Docker Images

# docker images

REPOSITORY          TAG                 IMAGE ID            CREATED             SIZE
docker.io/centos    6                   609c1f9b5406        7 weeks ago         194.5 MB

Gitlab Runner

Now, it is time to install and setup GitLab Runner.

In a nutshell this program, that is written in golang, will listen to every change on our repository and run every job that it can find on our yml file. But lets start with the installation:

# curl -s 'https://packages.gitlab.com/install/repositories/runner/gitlab-runner/config_file.repo?os=centos&dist=7&source=script' \
  -o /etc/yum.repos.d/gitlab-runner.repo

# yum -y install gitlab-runner

GitLab Runner Settings

We need to connect our project with the gitlab-runner.

 Project --> Settings --> CI/CD

or in our example:
http://192.168.122.131/root/libsodium/settings/ci_cd

click on the expand button on Runner’s settings and you should see something like this:

Register GitLab Runner

Type into your terminal:

# gitlab-runner register

following the instructions

[root@centos7 ~]# gitlab-runner register
Running in system-mode.                            

Please enter the gitlab-ci coordinator URL (e.g. https://gitlab.com/):
http://192.168.122.131/

Please enter the gitlab-ci token for this runner:
s6ASqkR8H9JysMSaFoLT

Please enter the gitlab-ci description for this runner:
[centos7]: 

Please enter the gitlab-ci tags for this runner (comma separated):

Whether to lock the Runner to current project [true/false]:
[true]: 

Registering runner... succeeded                     runner=s6ASqkR8

Please enter the executor: docker, ssh, virtualbox, docker-ssh+machine, kubernetes, docker-ssh, parallels, shell, docker+machine:
docker

Please enter the default Docker image (e.g. ruby:2.1):
centos:6

Runner registered successfully. Feel free to start it, but if it's running already the config should be automatically reloaded!
[root@centos7 ~]#

by refreshing the previous page we will see a new active runner on our project.

The Docker executor

We are ready to setup our first executor to our project. That means we are ready to run our first CI/CD example!

In gitlab this is super easy, just add a

New file --> Template --> gitlab-ci.yml --> based on bash

Dont forget to change the image from busybox:latest to centos:6

that will start a pipeline

GitLab Continuous Integration

Below is a gitlab ci test file that builds the rpm libsodium :

.gitlab-ci.yml

image: centos:6

before_script:
  - echo "Get the libsodium version and name from the rpm spec file"
  - export LIBSODIUM_VERS=$(egrep '^Version:' libsodium.spec | awk '{print $NF}')
  - export LIBSODIUM_NAME=$(egrep '^Name:'    libsodium.spec | awk '{print $NF}')

run-build:
  stage: build
  artifacts:
    untracked: true
  script:
    - echo "Install rpm-build package"
    - yum -y install rpm-build
    - echo "Install BuildRequires"
    - yum -y install gcc
    - echo "Create rpmbuild directories"
    - mkdir -p rpmbuild/{BUILD,RPMS,SOURCES,SPECS,SRPMS}
    - echo "Download source file from github"
    - curl -s -L https://github.com/jedisct1/$LIBSODIUM_NAME/releases/download/$LIBSODIUM_VERS/$LIBSODIUM_NAME-$LIBSODIUM_VERS.tar.gz -o rpmbuild/SOURCES/$LIBSODIUM_NAME-$LIBSODIUM_VERS.tar.gz
    - rpmbuild -D "_topdir `pwd`/rpmbuild" --clean -ba `pwd`/libsodium.spec

run-test:
  stage: test
  script:
    - echo "Test it, Just test it !"
    - yum -y install rpmbuild/RPMS/x86_64/$LIBSODIUM_NAME-$LIBSODIUM_VERS-*.rpm

run-deploy:
  stage: deploy
  script:
    - echo "Do your deploy here"

GitLab Artifacts

Before we continue I need to talk about artifacts

Artifacts is a list of files and directories that we produce at stage jobs and are not part of the git repository. We can pass those artifacts between stages, but you have to remember that gitlab can track files that only exist under the git-clone repository and not on the root fs of the docker image.

GitLab Continuous Delivery

We have successfully build an rpm file!! Time to deploy it to another machine. To do that, we need to add the secure shell private key to gitlab secret variables.

Project --> Settings --> CI/CD

stage: deploy

Lets re-write gitlab deployment state:

variables:
  DESTINATION_SERVER: '192.168.122.132'

run-deploy:
  stage: deploy
  script:
    - echo "Create ssh root directory"
    - mkdir -p ~/.ssh/ && chmod 700 ~/.ssh/

    - echo "Append secret variable to the ssh private key file"
    - echo -e "$SSH_PRIVATE_test_KEY" > ~/.ssh/id_rsa
    - chmod 0600 ~/.ssh/id_rsa

    - echo "Install SSH client"
    - yum -y install openssh-clients

    - echo "Secure Copy the libsodium rpm file to the destination server"
    - scp -o StrictHostKeyChecking=no rpmbuild/RPMS/x86_64/$LIBSODIUM_NAME-$LIBSODIUM_VERS-*.rpm  $DESTINATION_SERVER:/tmp/

    - echo "Install libsodium rpm file to the destination server"
    - ssh -o StrictHostKeyChecking=no $DESTINATION_SERVER yum -y install /tmp/$LIBSODIUM_NAME-$LIBSODIUM_VERS-*.rpm

and we can see that our pipeline has passed!

Possible Problems:

that will probable fail!

because our docker images don’t recognize gitlab.example.com.

Disclaimer: If you are using real fqdn - ip then you will probably not face this problem. I am referring to this issue, only for people who will follow this article step by step.

Easy fix:

# export -p EXTERNAL_URL="http://192.168.122.131" && yum -y reinstall gitlab-ce

GitLab Runner

GitLab Runner is not running !

# gitlab-runner verify
Running in system-mode.                            

Verifying runner... is alive                        runner=e9bbcf90
Verifying runner... is alive                        runner=77701bad

#  gitlab-runner status
gitlab-runner: Service is not running.

# gitlab-runner install  -u gitlab-runner -d /home/gitlab-runner/

# systemctl is-active gitlab-runner
inactive

# systemctl enable gitlab-runner
# systemctl start gitlab-runner

# systemctl is-active gitlab-runner
active

# systemctl | egrep gitlab-runner
  gitlab-runner.service     loaded active running   GitLab Runner

# gitlab-runner status
gitlab-runner: Service is running!

# ps -e fuwww | egrep -i gitlab-[r]unner
root      5116  0.4  0.1  63428 16968 ?        Ssl  07:44   0:00 /usr/bin/gitlab-runner run --working-directory /home/gitlab-runner/ --config /etc/gitlab-runner/config.toml --service gitlab-runner --syslog --user gitlab-runner