Deploy resources across a fleet

Connect to Git

Your Git repository contains the resources that you want to deploy across a fleet. To deploy those resources with a fleet package, you must connect your Git repository to Cloud Build.

Create your Git repository

This tutorial uses GitHub as the Git provider. To create a new GitHub repository, complete the following steps:

1. In your web browser, open GitHub.

2. If needed, sign in to your GitHub account. If you have access to other organizations or teams on GitHub, make sure that you're creating the repository with your personal account.

3. From the toolbar, select Add add and then click New repository.

4. Type fleet-package-tutorial as the name of the repository.

5. Keep Public selected as the repository visibility.

6. Select Create repository.

Connect the repository to Cloud Build

Config Sync's fleet package service uses Cloud Build to sync and fetch the Kubernetes resources from your Git repository.

To connect your GitHub repository to Cloud Build, complete the following steps:

1. Open the Cloud Build page in the Google Cloud console, and then select Repositories.

   Open the Repositories page

2. Ensure that you are on the 2nd gen Repositories page. If needed, select View repositories (2nd gen).

3. Click Create host connection.

4. In the Region menu, select us-central1 (Iowa) as your region.

5. In the Name field, type fleet-package-quickstart-connection as the name for your connection.

6. Click Connect.

7. If this is your first time connecting Cloud Build to your GitHub account, complete the following steps:

   1. Accept the request for your GitHub OAuth token. The token is stored in Secret Manager for use with Cloud Build GitHub Connection. Click Continue.
   2. Install Cloud Build into your GitHub repository. Select Install in a new account.
   3. In the new GitHub window that opens, select the GitHub account in which you created the fork of Cymbal Bank earlier. In a production environment, you might select other accounts or repositories that you have delegated access to.
   4. Follow any authentication prompts to confirm your identity in GitHub.
   5. In the GitHub window for Cloud Build repository access, choose Only select repositories.
   6. From the drop-down menu that lists repositories, select your fleet-package-tutorial repository.
   7. Click Save.

8. In the Cloud Build page in the Google Cloud console, click Link repository to connect a new Git repository to Cloud Build.

9. In the Connection menu, select fleet-package-quickstart-connection.

10. In the Repositories menu, select your fleet-package-tutorial repository.

11. Select Link.

Set up your fleet

In this section, you set up your fleet by creating clusters, registering them to a fleet, and installing Config Sync as a fleet package.

Create clusters

To demonstrate how you can use fleet packages to deploy resources across multiple clusters, this tutorial shows you how to create two clusters.

To create the two clusters and register them to your project's fleet:

1. Create a GKE cluster:

   gcloud container clusters create-auto cluster1 \
     --project=PROJECT_ID \
     --region=REGION \
     --fleet-project=PROJECT_ID \
     --release-channel=rapid
   

   Replace the following:

   • PROJECT_ID with your project ID.
   • REGION with the region that you want to create your cluster in, such as us-central1.

2. Create a second GKE cluster:

   gcloud container clusters create-auto cluster2 \
     --project=PROJECT_ID \
     --region=REGION \
     --fleet-project=PROJECT_ID \
     --release-channel=rapid
   

Install Config Sync as a fleet default

To use the fleet package service, Config Sync must be installed on both clusters. You can install Config Sync on multiple clusters at once, and any future clusters registered to the fleet, by completing the following steps:

1. In the Google Cloud console, go to the Feature Manager page.

   Go to Feature Manager

2. In the Config Sync pane, click Configure.

3. Click Customize fleet settings. In the dialog that appears, select Auto-upgrades. This setting ensures that your clusters have a Config Sync version that supports fleet packages.

4. Click Save changes.

5. Click Configure.

6. In the Configuring fleet settings confirmation dialog, click Confirm. If you haven't previously enabled Config Sync, clicking Confirm also enables the anthosconfigmanagement.googleapis.com API.

7. In the Clusters in the fleet table, select both clusters and then Click Sync to fleet settings. This installs Config Sync on both your Clusters with the settings that you configured.

   It can take a few minutes for the clusters to sync. You can proceed with the next steps when Config Sync shows as Installed.

Set up a service account for Cloud Build

To create the service account and grant the required permissions to Cloud Build, complete the following steps:

1. Create the service account:

   gcloud iam service-accounts create "quickstart-service-account"
   

2. Grant the service account permission to fetch resources from your Git repository by adding an IAM policy binding for the Resource Bundle Publisher role:

   gcloud projects add-iam-policy-binding PROJECT_ID \
      --member="serviceAccount:quickstart-service-account@PROJECT_ID.iam.gserviceaccount.com" \
      --role='roles/configdelivery.resourceBundlePublisher'
   

   If prompted, select None as the condition for the policy.

3. Grant the service account permission to write logs by adding an IAM policy binding for the Logs Writer role:

   gcloud projects add-iam-policy-binding PROJECT_ID \
      --member="serviceAccount:quickstart-service-account@PROJECT_ID.iam.gserviceaccount.com" \
      --role='roles/logging.logWriter'
   

   If prompted, select None as the condition for the policy.

Deploy resources across your fleet

In this tutorial, you add a Kubernetes manifest with an nginx deployment to your Git repository, publish a release, and then create a fleet package to deploy the nginx application.

Commit a Kubernetes manifest to your repository

To add your resources to GitHub and publish a release, complete the following steps:

1. In a web browser window of your GitHub repository, click Add file and then Create new file.

2. Name your file deployment.yaml and paste the following contents into it:

   apiVersion: apps/v1
   kind: Deployment
   metadata:
     name: nginx-deployment
   spec:
     replicas: 3
     selector:
       matchLabels:
         app: nginx
     template:
       metadata:
         labels:
           app: nginx
       spec:
         containers:
         - image: nginx:1.14.2
           name: nginx
           ports:
           - containerPort: 80
   

3. Click Commit changes...

4. In the confirmation dialog, keep Commit directly to the main branch selected and then click Commit changes.

5. On the main page of your repository, select Releases from the sidebar.

6. At the top of the page, choose Draft a new release.

7. Select the Choose a tag menu and type v1.0.0 as the tag. Click Create new tag.

8. Click Publish release.

Deploy a resource to clusters with a fleet package

To deploy the new resource, create a new fleet package:

1. This fleet package targets all of the clusters in your fleet since it doesn't contain a selector field. This also means any future clusters added to the fleet will have the nginx deployment automatically added.

   In your Cloud Shell, create a file named fleet-package.yaml with the following content:

   resourceBundleSelector:
     cloudBuildRepository:
       name: projects/PROJECT_ID/locations/us-central1/connections/fleet-package-quickstart-connection/repositories/REPOSITORY_NAME
       tag: v1.0.0
       serviceAccount: projects/PROJECT_ID/serviceAccounts/quickstart-service-account@PROJECT_ID.iam.gserviceaccount.com
       path:
   target:
     fleet:
       project: projects/PROJECT_ID
   rolloutStrategy:
     rolling:
       maxConcurrent: 1
   

   Replace REPOSITORY_NAME with the repository name from Cloud Build. This is usually in the format USERNAME-REPOSITORY_NAME.

2. Create the fleet package to start the rollout:

   gcloud alpha container fleet packages create fp-nginx.yaml \
      --source=fleet-package.yaml \
      --project=PROJECT_ID
   

3. Verify that the fleet package was created:

   gcloud alpha container fleet packages list
   

   You can click the link provided to view the streaming logs for the Cloud Build job.

   The fleet package starts rolling out the Kubernetes resources across your fleet.

4. In the Google Kubernetes Engine page of the Google Cloud console, go to the Workloads page to see an aggregated view of the workloads that are being deployed on all your GKE clusters:

   Open the Workloads page

   It can take a few minutes for the workloads to become available. You might also notice availability errors while Autopilot adjusts your resource requests for the new deployment.

   Notice how, because maxConcurrent: is set to 1 in your fleet package definition, the fleet package API waits until the nginx-deployment is fully deployed on one cluster before starting the deployment to the second cluster. If you changed the roll-out strategy to maxConcurrent: 2 or higher, the resources would deploy to both clusters simultaneously.

   After a few minutes, you will see two new workloads for the nginx-deployment on both your clusters. You might need to refresh the page.

You can continue to explore different deployment strategies with fleet packages. For example, you could add a new cluster to your fleet to observe that your workload gets automatically deployed to that new cluster. For more information about deployment strategies and variations, see Deploy fleet packages.