Blue-Green is a deployment model that uses two Docker hosts that you deploy and configure identically in a Kubernetes cluster. With the Blue and Green deployment model, you reduce the downtime that can occur in your environment when your pipelines in vRealize Automation Code Stream deploy your applications.

The Blue and Green instances in your deployment model each serve a different purpose. Only one instance at a time accepts the live traffic that deploys your application, and each instance accepts that traffic at specific times. The Blue instance receives the first version of your application, and the Green instance receives the second.

The load balancer in your Blue-Green environment determines which route the live traffic takes as it deploys your application. By using the Blue-Green model, your environment remains operational, users don't notice any downtime, and your pipeline continuously integrates and deploys your application to your production environment.

The pipeline that you create in vRealize Automation Code Stream represents your Blue-Green deployment model in two stages. One stage is for development, and the other stage is for production.

The vRealize Automation Code Stream pipeline workspace supports Docker and Kubernetes for continuous integration tasks and custom tasks.

For information about configuring the workspace, see Configuring the Pipeline Workspace.

Table 1. Development stage tasks for Blue-Green deployment
Task type Task
Kubernetes Create a namespace for your Blue-Green deployment.
Kubernetes Create a secret key for Docker Hub.
Kubernetes Create the service used to deploy the application.
Kubernetes Create the Blue deployment.
Poll Verify the Blue deployment.
Kubernetes Remove the namespace.
Table 2. Production stage tasks for Blue-Green deployment
Task type Task
Kubernetes Green gets the service details from Blue.
Kubernetes Get the details for the Green replica set.
Kubernetes Create the Green deployment, and use the secret key to pull the container image.
Kubernetes Update the service.
Poll Verify that the deployment succeeded on the production URL.
Kubernetes Finish the Blue deployment.
Kubernetes Remove the Blue deployment.

To deploy your application in your own Blue-Green deployment model, you create a pipeline in vRealize Automation Code Stream that includes two stages. The first stage includes the Blue tasks that deploy your application to the Blue instance, and the second stage includes Green tasks that deploy your application to the Green instance.

You can create your pipeline by using the CICD smart pipeline template. The template creates your pipeline stages and tasks for you, and includes the deployment selections.

If you create your pipeline manually, you must plan your pipeline stages. For an example, see Planning a CICD native build in vRealize Automation Code Stream before manually adding tasks.

In this example, you use the CICD smart pipeline template to create your Blue-Green pipeline.

Prerequisites

Procedure

  1. Click Pipelines > New Pipeline > Smart Templates > CI/CD template.
  2. Enter the information for the CI portion of the CICD smart pipeline template, and click Next.
  3. Complete the CD portion of the smart pipeline template
    1. Select the environments for your application deployment. For example, Dev and Prod.
    2. Select the service that the pipeline will use for the deployment.
    3. In the Deployment area, select the cluster endpoint for the Dev environment and the Prod environment.
    4. For the Production deployment model, select Blue-Green, and click Create.

    In the smart pipeline template, select the service, deployments, and click Blue-Green for the production deployment model.

Results

Congratulations! You used the smart pipeline template to create a pipeline that deploys your application to your Blue-Green instances in your Kubernetes production cluster on AWS.

Example: Example YAML code for some Blue-Green Deployment Tasks

The YAML code that appears in Kubernetes pipeline tasks for your Blue-Green deployment might resemble the following examples that create the Namespace, Service, and Deployment. If you need to download an image from a privately-owned repository, the YAML file must include a section with the Docker config Secret. See the CD portion of Planning a CICD native build in vRealize Automation Code Stream before using the smart pipeline template.

After the smart pipeline template creates your pipeline, you can modify the tasks as needed for your own deployment.

YAML code to create an example namespace:

apiVersion: v1
kind: Namespace
metadata:
  name: codestream-82855
  namespace: codestream-82855

YAML code to create an example service:

apiVersion: v1
kind: Service
metadata:
  labels:
    app: codestream-demo
  name: codestream-demo
  namespace: bluegreen-799584
spec:
  minReadySeconds: 0
  ports:
  - port: 80
  selector:
    app: codestream-demo
    tier: frontend
  type: LoadBalancer

YAML code to create an example deployment:

apiVersion: extensions/v1
kind: Deployment
metadata:
  labels:
    app: codestream-demo
  name: codestream-demo
  namespace: bluegreen-799584
spec:
  minReadySeconds: 0
  replicas: 1
  selector:
    matchLabels:
      app: codestream-demo
      tier: frontend
  template:
    metadata:
      labels:
        app: codestream-demo
        tier: frontend
    spec:
      containers:
      - image: ${input.image}:${input.tag}
        name: codestream-demo
        ports:
        - containerPort: 80
          name: codestream-demo
      imagePullSecrets:
      - name: jfrog-2
      minReadySeconds: 0

What to do next

To learn more about how you can use vRealize Automation Code Stream, see Tutorials for using vRealize Automation Code Stream.

To roll back a deployment, see How do I roll back my deployment in vRealize Automation Code Stream.

For additional references, see More resources for vRealize Automation Code Stream Administrators and Developers.