kubevirt

name: kubevirt description: KubeVirt VM operations via kubectl virt or oc virt. Use when the user wants to create, start, stop, console/VNC/SSH into, or inspect VMs on OpenShift/Kubernetes. Output CLI for the user to run — do not execute kubectl/oc yourself unless they ask.

For AI assistants

Do not run kubectl, oc, or virtctl in the user’s environment unless they explicitly ask you to execute commands. Give copy-paste bash blocks, explain placeholders (for example <vm-name>, <namespace>), and note prerequisites (logged-in cluster context, namespace, installed plugin).

KubeVirt CLI (`kubectl virt` / `oc virt`)

The kubectl virt plugin (OpenShift: oc virt) generates and manages KubeVirt VirtualMachine resources.

Install the plugin

Tell the user to install the virt plugin (pick one approach):

# Krew (recommended)
kubectl krew install virt

# Verify
kubectl virt version

On OpenShift, the same plugin is typically available as oc virt after installing OpenShift Virtualization.

Discover flags (user runs locally)

kubectl virt --help
kubectl virt create vm --help
kubectl virt ssh --help

Instance types and preferences (OpenShift)

Prefer instance types and preferences over raw --memory when the cluster exposes them:

kubectl get virtualmachineclusterinstancetype
kubectl get virtualmachineclusterpreference

Common patterns: u1.medium (universal), fedora / rhel.9 / windows.11.virtio preferences. With a DataSource that has annotations, --infer-instancetype and --infer-preference can select sizing automatically.

List and inspect VMs

kubectl get vm -A
kubectl get vm -n <namespace>
kubectl describe vm <vm-name> -n <namespace>
kubectl get vmi -n <namespace>

Start, stop, restart

kubectl virt start <vm-name> -n <namespace>
kubectl virt stop <vm-name> -n <namespace>
kubectl virt restart <vm-name> -n <namespace>

Create a VM (examples)

kubectl virt create vm prints YAML to stdout — the user pipes it to kubectl apply:

Registry containerdisk + instance type

kubectl virt create vm \
  --name=my-fedora \
  --instancetype=u1.medium \
  --preference=fedora \
  --volume-import=type:registry,url:docker://quay.io/containerdisks/fedora:latest,size:30Gi \
  --user=fedora \
  --ssh-key="$(cat ~/.ssh/id_rsa.pub)" \
  | kubectl apply -f -

Cluster DataSource with inferred sizing

kubectl virt create vm \
  --name=my-vm \
  --volume-import=type:ds,src:openshift-virtualization-os-images/fedora,size:30Gi \
  --infer-instancetype --infer-preference \
  --user=fedora \
  --ssh-key="$(cat ~/.ssh/id_rsa.pub)" \
  | kubectl apply -f -

Fallback when no instance types (memory only)

kubectl virt create vm \
  --name=my-vm \
  --memory=2Gi \
  --run-strategy=Always \
  --volume-import=type:registry,url:docker://quay.io/containerdisks/fedora:latest,size:30Gi \
  --user=fedora \
  --ssh-key="$(cat ~/.ssh/id_rsa.pub)" \
  | kubectl apply -f -

List DataSources (OpenShift golden images):

kubectl get datasource -n openshift-virtualization-os-images

Storage gotcha

If DataVolumes stay pending with missing access mode / storage class, the cluster may lack a default StorageClass. The user can mark one:

kubectl get storageclass
kubectl annotate storageclass <name> storageclass.kubernetes.io/is-default-class=true

Console and VNC

kubectl virt console <vm-name> -n <namespace>
kubectl virt vnc <vm-name> -n <namespace>

SSH (virtctl-style via plugin)

kubectl virt ssh fedora@<vm-name> -n <namespace>
kubectl virt ssh fedora@<vm-name> -n <namespace> --identity-file=~/.ssh/id_rsa
kubectl virt scp ./local.txt fedora@vmi/<vm-name>:./remote.txt -n <namespace>

Pause, migrate

kubectl virt pause vm <vm-name> -n <namespace>
kubectl virt unpause vm <vm-name> -n <namespace>
kubectl virt migrate <vm-name> -n <namespace>

When unsure of a subcommand, tell the user to run kubectl virt <command> --help locally.