Version: latest

Import and Deploy OVAs to Palette VMO

VMware vSphere Virtual Machines (VMs) can be exported using the Open Virtual Appliance (OVA) format. This allows users to share configurations and create templates for their VMs. These important templates need to be converted into Kubernetes compatible formats before they can be used in a Kubernetes cluster.

The Palette CLI provides the ability to import and deploy OVAs to Palette VMO.

Limitations

Only OVAs whose operating systems are included under virt-v2v supported guest systems can be imported.

Prerequisites

A Healthy VMO cluster. Refer to the Create a VMO Profile for further guidance.
- Ensure that your VMO cluster has sufficient capacity for any VMs that you deploy. Refer to the Environment Setup page for recommended environment sizes.

warning

If you need to provision Block storage volumes during the VM deployment process, add the following custom configuration to your VMO cluster OS pack. Applying this configuration may cause a cluster repave. For more information, refer to Repave Behaviors and Configurations

Additionally, we recommend provisioning volumes with the ReadWriteMany access mode to ensure that VMs can be live migrated.

kubeadmconfig:
  preKubeadmCommands:
    # Start containerd with new configuration
    - systemctl daemon-reload
    - systemctl restart containerd
  files:
    - targetPath: /etc/containerd/config.toml
  targetOwner: "root:root"
  targetPermissions: "0644"
  content: |
    ## template: jinja

    # Use config version 2 to enable new configuration fields.
    version = 2

    imports = ["/etc/containerd/conf.d/*.toml"]

    [plugins]
        [plugins."io.containerd.grpc.v1.cri"]
        sandbox_image = "registry.k8s.io/pause:3.9"
        device_ownership_from_security_context = true
        [plugins."io.containerd.grpc.v1.cri".containerd.runtimes.runc]
        runtime_type = "io.containerd.runc.v2"
        [plugins."io.containerd.grpc.v1.cri".containerd.runtimes.runc.options]
        SystemdCgroup = true

If you are in a proxied environment, you must configure the Containerized Data Importer (CDI) custom resource in order to deploy to a DataVolume. Refer to the CDI Configuration documentation.

If you have configured Nginx together with MetalLB, add an entry to /etc/hosts that maps the CDI upload proxy host name, for example cdi-uploadproxy.mycompany.io, to the Nginx load balancer’s public IP address.

Install kubectl command-line tool. Refer to the kubectl installation guide to learn more.
Install the virtctl command-line tool. Refer to the virtctl installation guide for more information.

Enablement

Download the Kubeconfig file of the VMO cluster.
Open a terminal window and set the environment variable KUBECONFIG to point to the file you downloaded.
```
export KUBECONFIG=<path-to-kubeconfig-file>
```

Execute the following command in your terminal. It creates a dedicated VM inside your VMO cluster which contains all the tools required for OVA import and deployment. The VM is named vmo-jh.

cat <<EOF | kubectl apply --filename -
apiVersion: kubevirt.io/v1
kind: VirtualMachine
metadata:
    name: vmo-jh
    namespace: default
spec:
dataVolumeTemplates:
- metadata:
         name: vmo-jh-dv
  spec:
    pvc:
        accessModes:
        - ReadWriteMany
        resources:
        requests:
            storage: 66Gi
    source:
         registry:
             url: gcr.io/spectro-images-public/release/vmo-jh:v4.5.0
running: true
template:
    metadata:
    labels:
        kubevirt.io/vm: vmo-jh
    spec:
        domain:
            cpu:
                cores: 2
                sockets: 2
                threads: 1
            devices:
                disks:
                - disk:
                    bus: virtio
                    name: datavolume-os
                - disk:
                    bus: virtio
                    name: cloudinitdisk
                interfaces:
                - masquerade: {}
                    model: virtio
                    name: default
            machine:
                type: q35
            resources:
                requests:
                    memory: 8Gi
        networks:
        - name: default
            pod: {}
        volumes:
        - dataVolume:
            name: vmo-jh-dv
            name: datavolume-os
        - cloudInitNoCloud:
            userData: |
                packages:
                - qemu-guest-agent
                runcmd:
                - ["sudo", "systemctl", "enable", "--now", "qemu-guest-agent"]
            name: cloudinitdisk
EOF

info

Depending on the configuration of your VMO cluster's CSI layer, you may need to change the configured ReadWriteMany access mode to ReadWriteOnce. Check the documentation for your chosen CSI to ensure you have the correct setting.

The VM provisioning and start up process will take a few minutes to complete. Execute the following command to check the status of your VMs.

kubectl get vms --all-namespaces

The vmo-jh VM appears in the state Running when it is ready for usage.

NAMESPACE              NAME                                 AGE     STATUS          READY
default                vmo-jh                               6m16s   Running         True

The VM is created with the Welcome@123 default password and is accessible through your cluster's kubeconfig file. You can secure access to the cluster by uploading a public key to the VM and adding it as an authorized key. You will then need to supply the corresponding private key when connecting to the VM.
```
virtctl scp public-key.pub root@vmo-jh.default:/root

virtctl ssh root@vmo-jh.default --local-ssh=true

mkdir -p /root/.ssh && cat /root/public-key.pub >> /root/.ssh/authorized_keys
```
info
If you are starting an SSH session from a Windows machine, then you should omit the --local-ssh=true flag and execute the command virtctl ssh root@vmo-jh.default only.
Copy your cluster's kubeconfig file to the VM. Replace the placeholders with local path to the kubeconfig file and the path to your private key. The kubeconfig file is required for subsequent operations.
```
virtctl scp <path-to-downloaded-kubeconfig-file> --identity-file <path-to-private-key> root@vmo-jh.default:/root
```
Download the newest Palette CLI binary. Copy the binary to the VM, in the same way you copied the kubeconfig file. Replace the placeholders with local path to the downloaded binary and the path to your private key.
```
virtctl scp <path-to-downloaded-palette-cli-binary> --identity-file <path-to-private-key> root@vmo-jh.default:/root
```

Start an SSH session with the VM.

virtctl ssh --identity-file <path-to-private-key> root@vmo-jh.default --local-ssh=true

Download an OVA file to import on the VM. Alternatively, you can upload the OVA file using virtctl scp as demonstrated in previous steps. This guide uses a publicly available Bitnami image for demonstration purposes.
```
curl --location https://downloads.bitnami.com/files/stacks/wordpress/6.2.2/bitnami-wordpress-6.2.2-r1-debian-11-amd64.ova --output bitnami-wordpress-6.2.2-r1-debian-11-amd64.ova
```
Set the environment variable KUBECONFIG to point to the file you uploaded. This allows your terminal session to connect to your VMO cluster.
```
export KUBECONFIG=<path-to-kubeconfig-file>
```

Execute the following command to start an interactive shell and begin the import process.

palette vmo import-ova

tip

You can save your configuration to a file, allowing you to revise your configuration and perform the import later.

palette vmo import-ova --config-only

The Palette CLI prompts you for information regarding the OVA you want to import.

Parameter	Description	Values
OVA Path	The path to the OVA you uploaded to your VM. The path for the example provided is `/root/bitnami-wordpress-6.2.2-r1-debian-11-amd64.ova/`.
Container Disk Upload Method	Indicate whether to upload the image directly to the target cluster as a `DataVolume` or build and push a Docker image. You will need to provide an existing image registry if you select Docker.	`DataVolume` / `Docker Image`
Kubeconfig Path	The path to the kubeconfig file you have uploaded to your VM.
DataVolume Namespace	The namespace to create your `DataVolume`, if you selected this option previously.
DataVolume Name	The name of your `DataVolume`.
Overhead Percentage for DataVolume Size	Set an overhead percentage for your `DataVolume` compared to the OVA specification. This parameter is optional and can be skipped with the value `-1`. If skipped, the filesystem overhead percentage will be inferred from the CDI Custom Resource in your VMO cluster. Refer to the CDI Configuration for further details.
Access Mode for the PVC	Set the access mode for your `DataVolume`. Ensure that your configured CSI supports your selection.	`ReadWriteMany` / `ReadWriteOnce`
Create a PVC with VolumeMode=Block	Indicate whether to set `Block` volume mode on the `DataVolume`.	`y` / `N`
StorageClass	The storage class on the destination that will be used to create the VM volume.
CDI Upload Proxy URL	Optionally provide a custom CDI upload proxy URL. If ingress is configured for the CDI upload proxy, the ingress hostname will be used by default and must be resolvable via DNS. If the CDI upload proxy is exposed via a NodePort, a node IP and ephemeral port will be used. Depending on how CDI and DNS are configured, you may need to edit `/etc/hosts` to ensure DNS resolution. You may also port-forward the CDI upload proxy via `kubectl --namespace cdi port-forward deployment/cdi-uploadproxy 8443` and provide `https://localhost:8443` as the CDI upload proxy URL. However, this approach will be less efficient.

The import may take a few minutes to complete. The Palette CLI outputs the path for your OVA configuration file. Make a note of it.
Your imported OVA is now ready to be deployed to a new VM inside your VMO cluster. Execute the following command to start an interactive shell and begin the deployment process. Replace the path placeholder with the path of your OVA configuration file.
```
palette vmo deploy-ova --config-file <path-to-ova-configuration-file>
```
The Palette CLI prompts you for information regarding the VM you want to deploy.

Parameter Description Values
VM Namespace The namespace to create your VM into.
VM Name The namespace of your VM.
Start VM Automatically? Indicate whether to start your VM after creation. y/ N
Depending on the configuration of your OVA, you may be prompted to provide additional values for the OVF template variables. The deployment of your VM should complete within a few minutes.

Parameter	Description	Values
VM Namespace	The namespace to create your VM into.
VM Name	The namespace of your VM.
Start VM Automatically?	Indicate whether to start your VM after creation.	`y`/ `N`

info

The import may generate a variety of warnings during the deployment process, including many repeated lines containing the word (BADINDEX). All of these warnings and messages can safely be ignored. If a fatal error is encountered, Palette CLI will exit completely.

You may need to make minor edits to the auto-generated VM specification if you want to configure Multus VLAN for each network interface. If uploading your image to a DataVolume upload fails, you may restart the upload after debugging any CDI issues. Execute the command palette vmo import-ova --config-file <path-to-ova-config.yaml> --skip-convert.

Validate

Log into Palette.
From the left Main Menu, select Clusters. Then, choose the VMO cluster that you deployed your OVA to. The Overview tab appears.
Select the Virtual Machines tab. Then, select your migration namespace from the Namespace drop-down Menu. Your deployed VM appears in the list of Virtual Machines. Select your VM to view its details. It is in a Running state.
Even though your VM is healthy, Palette displays two warnings. You can resolve them from the UI. Click Boot Order. The Virtual machine boot order dialog appears.
Select a boot source for your VM from the list of available disks. Then, click Save. The dialog closes.
Click the second warning button, named after the clusters Environment configuration. Select a new Serial Number for your environment ConfigMap. Then, click Save.
Select the Actions drop-down Menu. Then, select Restart to apply your changes. Once your changes have been applied, Palette will not display any warnings.

Limitations​

Prerequisites​

Enablement​

Validate​

Limitations

Prerequisites

Enablement

Validate