Understand KubeVirt basics to help get you started

Get to know KubeVirt basics

KubeVirt is the project that fills the gap between containers and virtualization by providing a platform where you can build, modify and deploy applications that reside in application containers and VMs. The advantage of this is KubeVirt can run containers as VMs on top of a Kubernetes infrastructure. It helps combine workloads where some parts run in VMs and other parts run in containers.

KubeVirt runs on top of Kubernetes, so you can use it in any common environment that's running Kubernetes. It does require the installation of some additional binaries, though. Virtctl is the most important because it offers control options for the VMs running on top of KubeVirt. Virtctl is the management utility you use after adding a VM to the configuration. It can manipulate the state of a container, and there is a virtctl console that enables you to connect to a console in the VM. You can manage KubeVirt VMs similarly to the way you manage traditional Kubernetes VMs.

Create KubeVirt containers and VMs using YAML files

You can create VMs the same way you create containers: using YAML manifest files. Below is a basic example of a YAML file that contains all of the necessary components:

apiVersion: kubevirt.io/v1alpha3

kind: VirtualMachine

metadata:

  name: testvm

spec:

  running: false

  template:

    metadata:

      labels:

        kubevirt.io/size: small

        kubevirt.io/domain: testvm

    spec:

      domain:

        devices:

          disks:

            - name: containerdisk

              disk:

                bus: virtio

            - name: cloudinitdisk

              disk:

                bus: virtio

          interfaces:

          - name: default

            bridge: {}

        resources:

          requests:

            memory: 64M

      networks:

      - name: default

        pod: {}

      volumes:

        - name: containerdisk

          containerDisk:

            image: kubevirt/cirros-registry-disk-demo

        - name: cloudinitdisk

          cloudInitNoCloud:

            userDataBase64: SGkuXG4=

Manage containers and VMs with KubeVirt

KubeVirt uses KVM as a back-end hypervisor. The KVM hypervisor integrates with standard Kubernetes components, such as networking and Persistent Volumes, which you can use to store files in a way that ensures survival after a reboot. This integration makes working with KubeVirt containers easy, especially in an environment already heavily oriented with Kubernetes.

KubeVirt also brings Kubernetes to VMs. For example, you can manage scalability using the VirtualMachineInstanceReplicaSet property, which is comparable to the standard Kubernetes Replica Set. Once you start a VM as a Replica Set, you can use a service to expose it. In this case, KubeVirt uses the service as a load balancer and directs the incoming traffic to a specific VM.

However, KubeVirt isn't ideal for the migration of large virtual environments to a Kubernetes platform. It doesn't have the tools to do so, and it also lacks broad hypervisor support to make that a plausible option. The project is still relatively young, but additional features are plausible.