How to install KubeOps Virtualization (kubevirt)

This guide describes the installation of the kubevirt virtualization software.
Tags:
Categories:

  5 minute read  

KubeVirt provides a powerful way to run virtual machines on top of Kubernetes, creating a unified platform for both containerized and traditional workloads. This guide walks through the essential steps for getting KubeVirt up and running using the KOSI package manager - from pulling and preparing the KubeVirt package, to creating a custom configuration file with your credentials, and finally verifying a successful installation in the kubevirt-namespace. Once these steps are complete, you’ll be ready to deploy and manage virtual machines alongside your existing container infrastructure - all within a single Kubernetes cluster.

KubeVirt is a specialized package that extends the capabilities of the KubeOps platform to support machine virtualization.

Requirements

To follow the next steps, make sure you have your credentials ready from your KubeOps.net account.

Make sure you have prepared the following points:

  • A KubeOps cluster with one master node.
  • Access to the master node with root privileges.

Step 1 - Download KubeOpsVM

  1. Log in to the master node remotely. For example, using ssh.
  2. If necessary, switch to your cluster user. If you forgot your cluster user, check your cluster-values.yaml.
su <your linux username>
  1. Log in to KOSI using your KOSI username.
kosi login -u [USERNAME]
  1. Search for available software packages by using the KOSI package manager. In this case, we restrict the search to KubeOps virtualization using kubeops-vm.
kosi search --hub kubeops-vm

  1. Log in to the local container registry on the master node using podman. Replace <master-ip> with the IP address of your master node.
podman login <master-ip>:30002 -u admin -p <harbor admin password> --tls-verify=false
  1. The package is named kubeops/kubevirt:2.2.0_Beta0 and lies on the hub kubeops-vm. Download the package using the kosi pull command and save it with the desired file name. In this example, the file is saved as kubevirt-2.2.0_Beta0.tgz. The -r flag sets the source registry the container images are pulled from, and -t sets the local target registry the images are pushed to.
kosi pull --hub kubeops-vm kubeops/kubevirt:2.2.0_Beta0 -o kubevirt-2.2.0_Beta0.tgz -r <master-ip>:30002/kubeops -t 127.0.0.1:30002/kubeops

Step 2 - Remove the Taint of the Master Node (only necessary on single node cluster)

This step is only required for a single-node cluster.

If your cluster consists of only one node, you must remove the taint from the master node so that workloads can be scheduled on it.

Replace [NODENAME] with the name of you master node.

kubectl taint nodes [NODENAME] node-role.kubernetes.io/control-plane:NoSchedule-

Step 3 - Create Kubevirt Configuration File

Before installing the package, create a YAML configuration file. Save the following content in the same directory as the package file and name the file kubevirt-values.yaml.

Replace [your username] and [your harbor cli secret] with your KOSI username and your Harbor CLI secret.

The harbor-values must be adapted to your harbor instance.

apiVersion: kubeops/kubeopsctl/enterprise/beta/v1

deleteNs: false
localRegistry: true

packages:
  - name: kubevirt-manager
    enabled: true
    values:
      standard:
        namespace: kubevirt
        hostname: kubeops-vms.local
        nodeport: 30733
        ingressName: nginx
      advanced: {}

  - name: kubevirt
    enabled: true
    hub: kubeops-vm
    values:
      standard:
        namespace: kubevirt

        imagePullSecretCred:
          registry: 127.0.0.1:30002/kubeops
          user: [your username]
          password: [your harbor cli secret] # must be CLI Token from harbor

        kubevirtobject:
          spec:
            configuration:
              developerConfiguration:
                logVerbosity:
                  virtLauncher: 2
                  virtHandler: 3
                  virtController: 4
                  virtAPI: 5
                  virtOperator: 6

                featureGates:
                  - ExpandDisks
                  - CPUManager
                  - ExperimentalIgnitionSupport
                  - HypervStrictCheck
                  - Sidecar
                  - HostDevices
                  - Snapshot
                  - VMExport
                  - HotplugVolumes
                  - HostDisk
                  - ExperimentalVirtiofsSupport
                  - DownwardMetricsFeatureGate
                  - Root
                  - ClusterProfiler
                  - WorkloadEncryptionSEV
                  - VSOCK
                  - DisableCustomSELinuxPolicy
                  - KubevirtSeccompProfile
                  - DisableMDEVConfiguration
                  - PersistentReservation
                  - VMPersistentState
                  - MultiArchitecture
                  - AutoResourceLimitsGate
                  - AlignCPUs
                  - VolumesUpdateStrategy
                  - VolumeMigration
                  - InstancetypeReferencePolicy

                useEmulation: false

        cdiobject:
          spec:
            config:
              insecureRegistries:
                - registry.preprod.kubeops.net
              scratchSpaceStorageClass: rook-cephfs # Optional; Default: rook-cephfs
  - name: harbor
    enabled: false
    values:
      standard:
        namespace: harbor # optional, default is harbor
        ingressClassName: nginx # optional, default is nginx; set to "traefik" to route via Traefik
        harborpass: "password" # required: set password for harbor access
        databasePassword: "Postgres_Password" # required: set password for database access
        redisPassword: "Redis_Password" # required: set password for redis access
        externalURL: http://10.2.10.11:30002 # required, the ip address and port, from which harbor is accessable outside of the cluster
        nodePort: 30002 # required
        hostname: harbor.local # required
        harborPersistence:
          persistentVolumeClaim:
            registry:
              size: 40Gi # optional, default is 40Gi
              storageClass: "rook-cephfs" #optional, default is rook-cephfs
            jobservice:
              jobLog:
                size: 1Gi # optional, default is 1Gi
                storageClass: "rook-cephfs" #optional, default is rook-cephfs
            database:
              size: 1Gi # optional, default is 1Gi
              storageClass: "rook-cephfs" #optional, default is rook-cephfs
            redis:
              size: 1Gi # optional, default is 1Gi
              storageClass: "rook-cephfs" #optional, default is rook-cephfs
            trivy: 
              size: 5Gi # optional, default is 5Gi
              storageClass: "rook-cephfs" #optional, default is rook-cephfs
      advanced: #  All values from https://artifacthub.io/packages/helm/harbor/harbor/1.8.1#configuration are overwritable

Step 4 - Install KubeOps-VM

  1. Install the KubeVirt package using the prepared configuration file you created.
kosi install -p kubevirt-2.2.0_Beta0.tgz -f kubevirt-values.yaml
  1. Verify the installation by checking the pods in the kubevirt namespace.
kubectl get pods -n kubevirt

The installation is complete when the status of all pods show the status “Running”. For example:

the following example, the kubevirt namespace is used. If KubeVirt is deployed in a different namespace, replace the value passed to the -n parameter.

kubectl -n kubevirt create secret docker-registry kubeops-vm --docker-server=127.0.0.1:30002 --docker-username=admin --docker-password=<admin password - see harborpass value in kubeopsctl.yaml>

Step 5 - Pull and Install KubeVirt Manager Package

  1. Pull the KubeVirt Manager package.
kosi pull --hub kubeops kubeops/kubevirt-manager:2.2.0_Beta0 -o kubevirt-manager-2.2.0_Beta0.tgz -r master-ip:30002/kubeops -t 127.0.0.1:30002/kubeops
  1. Install the KubeVirt Manager package.
kosi install -p kubevirt-manager-2.2.0_Beta0.tgz -f kubevirt-values.yaml

Next Steps

After the installation is complete, you can start creating and deploying a custom virtual machine. For more information, see How to create and deploy a Custom Virtual Machine using KubeOps.