Edit

Share via

Create deployments

  • Article

Applies to: AKS on Azure Stack HCI 22H2, AKS on Windows Server

This article provides an overview of Kubernetes deployments, and describes how to create and update these deployments in AKS enabled by Azure Arc.

Kubernetes deployments overview

A deployment refers to a Kubernetes object that manages the performance and specifies the desired behavior of a pod. It specifies the application's life cycle, including the pods assigned to the application. A deployment provides a way to communicate your desired state for your application, and the controller works on changing the present state into your desired state.

Deployments automate the process to launch pod instances and ensure they run as defined across all nodes within the cluster. Administrators and IT professionals use deployments to communicate what they want from an application, and then Kubernetes takes all the necessary steps to create the desired state of the application.

While deployments define how your applications run, they do not guarantee where your applications are located within your cluster. For example, if your application requires an instance of a pod on every node, you want to use a DaemonSet. For stateful applications, a StatefulSet provides unique network identifiers, persistent storage, and ordered deployment/scaling.

The Kubernetes deployment object lets you:

  • Deploy a replica set or a pod.
  • Scale the number of instances of an application up or down.
  • Update every running instance of an application.
  • Roll back all running instances of an application to another version.
  • Pause or continue a deployment.

For more information, see Kubernetes Deployments.

Create a deployment

To create a deployment, you can use the kubectl apply or kubectl create commands. Since the required number of pods is maintained and monitored, they're running and available after the deployment is created. If a pod fails, Kubernetes immediately rolls out a replica of the pod to take its place in the cluster.

The following example describes the features of a deployment manifest file in YAML format:

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

To view the deployment, the replica set, and the pods, run the following command:

kubectl get deployment, replicaset, pod

Update a deployment

The main advantage of deployments is the ability to automatically update your Kubernetes program. Without deployments, you would have to manually end all old pods, start new pod versions, and run a check to see if there are any problems when creating pods. You can run kubectl describe deployment to see the order in which the pods were brought up and removed.

Deployments automate the update process, as you simply update the pod template or the desired state. The deployment alters the program state in the background with actions, such as creating new pods or allocating more resources, until the chosen update is in place.

If there are problems in the deployment, Kubernetes automatically rolls back to the previous version. You can also explicitly roll back to a specific version using the kubectl rollout undo command, or you can use the kubectl rollout pause to temporarily halt a deployment.

Strategies for updating deployments

Kubernetes provides several deployment strategies so you can update in various ways to suit the needs of your environment. The three most common update strategies are:

  • Rolling update: This update is a gradual process that allows you to update your Kubernetes system with only a minor effect on performance and no downtime. It minimizes downtime at the cost of update speed.
  • Recreation: This strategy is an all-or-nothing process that allows you to update all aspects of the system at once with a brief downtime period. It updates quickly but causes downtime.
  • Canary: This strategy is a partial update process that allows you to test your new program version on real users without a commitment to a full rollout. It quickly updates for a select few users with a full rollout later.

Next steps