GitLabTech&Tips

Google Kubernetes Engine에 GitLab CI/CD를 15분 이내에 사용 가능

Source: GitLab Blog | Author: Elliot Rushton

설치형 GitLab을 사용하는 경우, GitLab의 Google Kubernetes Engine(GKE)과의 연동 기능을 사용하여 GitLab CI를 시작하는 것은 몇 가지 간단한 단계로 이루어질 수 있습니다. 우리는 Kubernetes 클러스터와 관련된 세부 설정 지침을 제공하는 몇 개의 블로그 포스트와 문서를 가지고 있습니다. 이 게시물에서는 GitLab CI/CD를 15분 이내에 사용할 수 있도록 필수 단계를 강조합니다.

By using the GitLab and GKE integration, with one click, you install GitLab Runners on GKE and immediately start running your CI pipelines. Runners are the lightweight agents that execute the CI jobs in your GitLab CI/CD pipeline.

Prerequisites:

The following pre-requisities will need to have been configured in order for you to use the built in GitLab GKE integration:

  • GitLab instance installed and configured with user credentials
  • Google OAuth2 OmniAuth Provider installed and configured on your GitLab instance
  • A Google Cloud project with the following APIs enabled:
    • Google Kubernetes Engine API
    • Cloud Resource Manager API
    • Cloud Billing API

Get started

Setup pipeline

Step 1

We’re going to add a shared runner at the instance level. First, as an administrator, click the “Admin Area” icon

Runner setup step 1

Then on the left menu, select “Kubernetes”

Runner setup step 2

Step 2

Click the green “Add Kubernetes cluster” button.

Runner setup step 3

Step 3

The screen to “Add a Kubernetes cluster integration” should come up. Click on the “Google GKE” icon on the right.

Runner setup step 4

Step 4

Give your cluster a name, and select a “Google Cloud Platform project” from your linked GCP account. If no projects are populated in the menu then either your Google OAUTH2 integration isn’t configured correctly or your project is missing the needed permissions. Check that these are set up and that the APIs mentioned in the prerequisites above are enabled.

Choose a zone in which to run your cluster. For the purposes of running CI, the number of nodes in your cluster is going to be how many simultaneous jobs you can run at given time. As we are using the built-in GitLab Google Kubernetes integration, you can set a maximum of four nodes. Here we set that to three.

Click “Create Kubernetes Cluster”

Runner setup step 5

It takes a few minutes for the cluster to be created. While it’s happening you should see a screen like this. You can leave this screen and come back (by going to “Admin Area> Kubernetes > [your cluster name]”)

Runner setup step 6

Step 5

Once the cluster has been created, we need to install two applications. First, install “Helm Tiller” by clicking on the “Install” button next to it.

Runner setup step 7

This takes a moment, but should be much quicker than creating the cluster initially was.

Runner setup step 8

Step 6

Now that Helm Tiller is installed, more applications can be installed. For this tutorial we only need to install the “GitLab Runner” application. Click the install button next to GitLab Runner.

Runner setup step 9

Again, this should go pretty quickly.

Runner setup step 10

Once done, the button will change to an “Uninstall” button. You’re now set up with shared runners on your GitLab instance and can run your first CI pipeline!

Runner setup step 11

Next steps

Now that you are up and running with GitLab CI/CD on GKE, you can build and run your first GitLab CI/CD pipeline. Here are links to a few resources to get you started.

If you are planning to manage your own fleet of GitLab Runners, then you may also be thinking about how best to set up autoscaling of GitLab Runners. As we have just set up your first Runner on GKE, then you can review the GitLab Runner Kubernetes Executor docs for additional details as to how the GitLab Runner uses Kubernetes to run builds on a Kubernetes cluster.

Other resources

Cover image by Agê Barros on Unsplash

댓글 남기기