Version: latest

Create Disk Templates

Palette Virtual Machine Orchestrator (VMO) enables you to quickly deploy Virtual Machines (VMs) by creating VM templates.

When deployed from a template, VMs include a disk template containing a pre-installed copy of the desired Operating System (OS). The Kubevirt Containerized Data Importer (CDI) imports templates through Data Volumes. This guide demonstrates how to implement your own disk and VM templates using Kubevirt Data Volumes.

Prerequisites

A VMO profile. Check out the Create a VMO Profile guide to learn how you can create this profile.
A cluster deployed with this VMO profile. Check out the Deploy a Cluster tutorial for detailed steps on how you can deploy clusters to a public cloud.

Create a Template

Create a new Add-on Profile with the following manifest. Check out the Add a Manifest guide for more information.

The provided manifest defines a DataVolume that imports the example disk template for Ubuntu 22.04 into the vmo-golden-images namespace. This snippet has three important configuration options.
- cdi.kubevirt.io/storage.deleteAfterCompletion that prevents deleting the DataVolume object after the import process completes. This annotation is necessary to stop Palette from continuously recreating the DataVolume object.
- volumeMode that aligns with the mode that CSI supports for ReadWriteMany access. For some CSIs, such as Rook-Ceph and EMC PowerMax, it's the Block mode, while for other CSIs, such as Portworx, Longhorn, and EMC PowerFlex, it's the Filesystem mode. Make sure to check the volumeMode setting if you choose another CSI.
- registry that specifies a container disk on an external Docker registry from which data should be imported. Check the CDI DataVolumes documentation for other possible source values.
```
apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: "template-ubuntu-2204"
  namespace: "vmo-golden-images"
  annotations:
    cdi.kubevirt.io/storage.deleteAfterCompletion: "false"
spec:
  storage:
    accessModes:
      - ReadWriteMany
    resources:
      requests:
        storage: 50Gi
    volumeMode: Block # or "Filesystem", depending on the CSI used
  source:
    registry:
      url: "docker://gcr.io/spectro-images-public/release/vm-dashboard/os/ubuntu-container-disk:22.04"
```
warning
When the Filesystem mode is selected, the CDI will automatically apply a small overhead factor to the Persistent Volume Claim (PVC) size to account for the storage space lost due to the filesystem overhead. As a result, the PVC for such a template will be slightly larger. Any clones made from this PVC will also use this larger minimum size.

Create another manifest in this cluster profile and add the following code snippet.

This snippet has three notable configurations.

spec.dataVolumeTemplates.spec that defines which source PVC to clone from and what PVC to clone into.
spec.dataVolumeTemplates.spec.source that points to the template-ubuntu-2204 PVC in the vmo-golden-images namespace.
spec.dataVolumeTemplates.spec.pvc that creates a new target PVC with the specified accessMode and volumeMode characteristics. The size of the new PVC must be the same or larger as the source PVC.

warning

To enable CSI-assisted cloning, you must fulfill a few prerequisites. For instance, the source and target PVCs must share the same Storage Class and Volume Mode. The CDI will automatically fall back to slower host-assisted cloning if you don't meet these requirements.

apiVersion: spectrocloud.com/v1
kind: VmTemplate
metadata:
  name: ubuntu-2204
spec:
  dataVolumeTemplates:
    - metadata:
        name: ubuntu-2204
      spec:
        source:
          pvc:
            name: template-ubuntu-2204
            namespace: vmo-golden-images
        pvc:
          accessModes:
            - ReadWriteMany
          resources:
            requests:
              storage: 50Gi
          volumeMode: Block # or "Filesystem", depending on the CSI used
  template:
    metadata:
      annotations:
        descheduler.alpha.kubernetes.io/evict: "true"
    spec:
      domain:
        cpu:
          cores: 2
          sockets: 1
          threads: 1
        devices:
          disks:
            - disk:
                bus: virtio
              name: datavolume-os
            - disk:
                bus: virtio
              name: cloudinitdisk
          interfaces:
            - masquerade: {}
              name: default
              model: virtio
        machine:
          type: q35
        resources:
          limits:
            memory: 2Gi
          requests:
            memory: 2Gi
      networks:
        - name: default
          pod: {}
      volumes:
        - dataVolume:
            name: ubuntu-2204
          name: datavolume-os
        - cloudInitNoCloud:
            userData: |
              #cloud-config
              ssh_pwauth: True
              chpasswd: { expire: False }
              password: spectro
              disable_root: false
              runcmd:
                - apt-get update
                - apt-get install -y qemu-guest-agent
                - systemctl start qemu-guest-agent
          name: cloudinitdisk

tip

To enable CSI-assisted cloning into larger PVCs than the source PVC, set allowVolumeExpansion : true on your defined Storage Class.

When the CDI clones a PVC, it runs under the default service account in the namespace of the target PVC. When the source PVC is in a different namespace, you must give the required permissions to the service account. The VMO pack version 4.2.0 (or higher) does this automatically through its default pack specification. This configuration uses the vmEnabledNamespaces option to specify the namespaces for which the permissions are configured.
```
charts:
  virtual-machine-orchestrator:
    # This namespace will be used to store golden images
    goldenImagesNamespace: "vmo-golden-images"

    # These namespaces will be created and set up to deploy VMs into
    vmEnabledNamespaces:
      - "default"
      - "virtual-machines"
```

By default, the CDI contains an internal database of CSIs, with the CSI supported access modes, volume modes, and CSI-assisted cloning support. However, we recommend creating a StorageProfile to adjust these parameters to your desired values.

Create another manifest in this cluster profile and add the following code snippet.

This snippet has three notable configurations that define the storage class.

The storage class has the name ceph-block.
The storage class supports the Block volume mode with the ReadWriteMany access mode.
The storage class uses csi-clone for cloning operations.

apiVersion: cdi.kubevirt.io/v1beta1
kind: StorageProfile
metadata:
  name: ceph-block
spec:
  claimPropertySets:
    - accessModes:
        - ReadWriteMany
      volumeMode: Block
  cloneStrategy: csi-clone

Expand the following section to learn how you can use storage profiles for CSI-assisted cloning.

Storage Profiles in VM Templates

If Storage Profiles are not manually defined, the CDI will automatically generate one for every Storage Class based on the information in its internal database. We recommend verifying that these parameters are accurate for your use case and, if not, overwriting them to define a Storage Profile manually.

Per the CDI documentation, you can optionally replace the target pvc sections in your vmTemplate resources with a storage section once you have defined a storage profile.

Using storage instead of PVC will automatically discover default values for accessMode, volumeMode and, optionally, even the size from the source PVC and its associated storage profile.

A benefit of storage over pvc is that the CDI will automatically calculate the size overhead for Filesystem mode storage classes, allowing the VM template to use the original disk size instead of the one adjusted for the filesystem overhead.

If you want to use this template, replace the contents of the manifest you added in step two with the following snippet.

apiVersion: spectrocloud.com/v1
kind: VmTemplate
metadata:
  name: ubuntu-2204
spec:
  dataVolumeTemplates:
    - metadata:
        name: ubuntu-2204
      spec:
        source:
          pvc:
            name: template-ubuntu-2204
            namespace: vmo-golden-images
        storage:
          resources:
            requests:
              storage: 50Gi
  template:
    metadata:
      annotations:
        descheduler.alpha.kubernetes.io/evict: "true"
    spec:
      domain:
        cpu:
          cores: 2
          sockets: 1
          threads: 1
        devices:
          disks:
            - disk:
                bus: virtio
              name: datavolume-os
            - disk:
                bus: virtio
              name: cloudinitdisk
          interfaces:
            - masquerade: {}
              name: default
              model: virtio
        machine:
          type: q35
        resources:
          limits:
            memory: 2Gi
          requests:
            memory: 2Gi
      networks:
        - name: default
          pod: {}
      volumes:
        - dataVolume:
            name: ubuntu-2204
          name: datavolume-os
        - cloudInitNoCloud:
            userData: |
              #cloud-config
              ssh_pwauth: True
              chpasswd: { expire: False }
              password: spectro
              disable_root: false
              runcmd:
                - apt-get update
                - apt-get install -y qemu-guest-agent
                - systemctl start qemu-guest-agent
          name: cloudinitdisk

Click on Confirm & Create.
Complete the cluster profile configuration and click on Finish Configuration.

Validate

Navigate to the left Main Menu and select Clusters.
Select your cluster and click the Profile tab.
Add the newly created Add-on Profile to your cluster profile. Click on Save.

Your cluster applies the changes. As soon as the cluster finishes updating, the CDI begins to import the data into the target PVC in the vmo-golden-images namespace. The target PVC has the same name as the DataVolume object.
Once the cluster updates, this VM template is available to the VMs you create on your cluster. Check out the Deploy VM From a Template guide for more information.

Resources

To learn more about the Palette VMO, we encourage you to check out the reference resources below.

Prerequisites​

Create a Template​

Validate​

Resources​

Prerequisites

Create a Template

Validate

Resources