What is a Deployment ?

A Deployment is used to manage one or more Pods with the same specification. It manages the Pod lifecycle:

• creation and deletion
• scaling
• progressive deployment (Rollout) and rollback

There are an additional level of abstraction:

• Deployment: creates and updates applications
• ReplicaSet: ensures the number of Pod replicas
• Pod: smallest workload unit containing one or more containers

From Pod to Deployment

Pod are rarely created directly, they are usually managed by higher level resources such as:

• Deployment
• DaemonSet
• Job
• CronJob
• StatefulSet

The choice of resource depends on the application’s use case. For example, a Deployment manages the lifecycle of multiple identical Pods and ensures that a specified number of Pods are always running.

The following specification defines a Pod with a unique container running the nginx:1.20 image:

apiVersion: v1
kind: Pod
metadata:
  name: www
  labels:             
    app: www
spec:
  containers:
  - name: www
    image: nginx:1.20
    ports:
    - containerPort: 80

If we want to manage 3 identical Pods having the previous specification, we can specify a Deployment as follows:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: www
spec:
  replicas: 3
  selector:
    matchLabels:
      app: www
  template:
    metadata:
      labels:
        app: www
    spec:
      containers:
      - name: www
        image: nginx:1.20
        ports:
        - containerPort: 80

We can create this Deployment:

kubectl apply -f deploy.yaml

If a Pod is deleted, a new one is automatically created to replace it, ensuring that 3 Pods (replicas) are always running as specified

If we need to update an application, in order to use a newer image for the containers, we can easily do so modifying the Deployment specification and applying the new version.

Below is the updated deploy.yaml specification, the nginx image was changed from nginx:1.20 to nginx:1.22:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: www
spec:
  replicas: 3
  selector:
    matchLabels:
      app: www
  template:
    metadata:
      labels:
        app: www
    spec:
      containers:
      - name: www
        image: nginx:1.22
        ports:
        - containerPort: 80

We can send this new version of the specification to Kubernetes:

kubectl apply -f deploy.yaml

The Pods running nginx:1.20 will gradually be replaced with Pods running nginx:1.22

Base commands

• Creating a Deployment from a YAML file

kubectl apply -f deploy.yaml

• Creating a Deployment using an imperative command

kubectl create deploy www --image=nginx:1.22 --replicas=3

• Listing existing Deployments

kubectl get deploy

• Getting the Deployment’s properties

kubectl get deploy www -o yaml

• Getting the Deployment’s main properties

kubectl describe deploy www

• Changing the number of Pods managed by a Deployment (scaling operation)

kubectl scale deploy www --replicas=5

• Removing a Deployment

kubectl delete deploy www

Using Deployments in the VotingApp

The previous exercise left the app with the following components:

We will now use Deployments to manage Pods instead of running the Pods directly. The Pods created by these Deployments will be exposed by the same Service.