Interconnecting GitLab with SWR and CCE for CI/CD

Background

GitLab is an open-source version management system developed with Ruby on Rails for Git project repository management. It supports web-based access to public and private projects. Similar to GitHub, GitLab allows you to browse source code, manage bugs and comments, and control team member access to repositories. You will find it very easy to view committed versions and file history database. Team members can communicate with each other using the built-in chat program (Wall).

GitLab provides powerful CI/CD functions and is widely used in software development.

Figure 1 GitLab CI/CD process

This section describes how to interconnect GitLab with SWR and CCE for CI/CD.

Preparations

  1. Create a CCE cluster and a node and bind an EIP to the node for downloading an image during GitLab Runner installation.
  2. Download and configure kubectl to connect to the cluster.
  3. Install Helm 3.

Installing GitLab Runner

Log in to GitLab, choose Settings > CI/CD in the project view, click Expand next to Runners, and search for the GitLab Runner registration URL and token.

Create the values.yaml file and fill in the following information:

# Registration URL
gitlabUrl: https://gitlab.com/
# Registration token
runnerRegistrationToken: "******"
rbac:
    create: true
runners:
    privileged: true

Create a GitLab namespace.

kubectl create namespace gitlab

Install GitLab Runner using Helm.

helm repo add gitlab https://charts.gitlab.io 
helm install --namespace gitlab gitlab-runner -f values.yaml gitlab/gitlab-runner --version=0.43.1

After the installation, you can obtain the gitlab-runner workload on the CCE console and view the connection information in GitLab later.

Creating an Application

Place the application to be created in the GitLab project repository. This section takes Nginx modification as an example. For details, visit https://gitlab.com/c8147/cidemo/-/tree/main.

The following files are included:

The preceding files are only examples. You can replace or modify them accordingly.

Configuring Global Variables

When using pipelines, build an image, upload it to SWR, and run kubectl commands to deploy the image in the cluster. Before performing these operations, you must log in to SWR and obtain the credential for connecting to the cluster. You can define the information as variables in GitLab.

Log in to GitLab, choose Settings > CI/CD in the project view, and click Expand next to Variables to add variables.

Creating a Pipeline

Log in to Gitlab and add the .gitlab-ci.yml file to Repository.

The content is as follows:

# Define pipeline stages, including package, build, and deploy.
stages:
  - package  
  - build
  - deploy
# If no image is specified in each stage, the default image docker:latest is used.
image: docker:latest
# In the package stage, only printing is performed.
package:
  stage: package
  script:
    - echo "package"
# In the build stage, the Docker-in-Docker mode is used.
build:
  stage: build
  # Define environment variables for the build stage.
  variables:
    DOCKER_HOST: tcp://docker:2375
  # Define the image for running Docker-in-Docker.
  services:
    - docker:18.09-dind
  script:
    - echo "build"
    # Log in to SWR.
    - docker login -u $project@$swr_ak -p $swr_sk swr.eu-de.otc.t-systems.com
    # Build an image. k8s-dev is the organization name in SWR. Replace it to the actual name.
    - docker build -t swr.eu-de.otc.t-systems.com/k8s-dev/nginx:$CI_PIPELINE_ID .
    # Push the image to SWR.
    - docker push swr.eu-de.otc.t-systems.com/k8s-dev/nginx:$CI_PIPELINE_ID
deploy:
  # Use the kubectl image.
  image: 
    name: bitnami/kubectl:latest
    entrypoint: [""]
  stage: deploy
  script:
    # Configure the kubeconfig file.
    - mkdir -p $HOME/.kube
    - export KUBECONFIG=$HOME/.kube/config
    - echo $kube_config |base64 -d > $KUBECONFIG
    # Replace the image in the k8s.yaml file.
    - sed -i "s/<IMAGE_NAME>/swr.eu-de.otc.t-systems.com\/k8s-dev\/nginx:$CI_PIPELINE_ID/g" k8s.yaml
    - cat k8s.yaml
    # Deploy an application.
    - kubectl apply -f k8s.yaml

After the .gitlab-ci.yml file is saved, the pipeline is started immediately. You can view the pipeline execution status in GitLab.

Verifying Deployment

After the pipeline is deployed, locate the nginx-test Service on the CCE console, query its access address, and run the curl command to access the Service.

# curl xxx.xxx.xxx.xxx:31111
Hello Gitlab!

If the preceding information is displayed, the deployment is correct.

Common Issues

Parent topic: DevOps