Import a virtual machine image using the `DataVolume` API

You can import a virtual machine (VM) image into your OpenShift Virtualization cluster by using the DataVolume API. A DataVolume automates PVC creation and image data transfer from an external source by using the Containerized Data Importer (CDI).

Supported import sources

The DataVolume API supports importing VM images from several external sources. The Containerized Data Importer (CDI) detects the image format and converts it as needed.

The following import sources are available:

HTTP or HTTPS: You specify the URL of the image file by using the source.http field. CDI downloads the image directly from the web server. You can optionally supply TLS certificates for custom certificate authorities and credentials for Basic authentication.
Container registry: You specify the image location by using the source.registry field with a docker:// URL prefix. CDI pulls the image from the registry by using standard container image pull mechanisms. You can supply registry credentials in a Secret object if the registry requires authentication.
S3 object storage: You specify the object URL by using the source.s3 field. CDI downloads the image from an S3-compatible storage service. You must supply access credentials in a Secret object.
oVirt (ImageIO): You specify the connection details by using the source.imageio field. CDI connects to the oVirt ImageIO API to transfer disk images from a Red Hat Virtualization environment.
VMware (VDDK): You specify the connection details by using the source.vddk field. CDI uses the VMware Virtual Disk Development Kit (VDDK) to transfer disk images from a VMware vSphere environment.

Supported image formats include QCOW2, raw, ISO, and compressed archive files.

Import a VM image from a container registry

You can import a virtual machine (VM) image from a container registry into a persistent volume claim (PVC) by using the DataVolume API. The Containerized Data Importer (CDI) creates a data volume and an associated PVC to store the imported image.

Prerequisites

You have access to the container registry.
If the registry is private, registry credentials are stored in a Secret object.
You have installed the OpenShift CLI (oc).

Procedure

Create a DataVolume object that specifies the container registry as the source:
```
apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: registry-image-datavolume
spec:
  source:
    registry:
      url: "docker://registry.example.com/my-vm-image:latest"
      pullMethod: node
      secretRef: <registry_secret>
      certConfigMap: <ca_certs_configmap>
  storage:
    resources:
      requests:
        storage: 20Gi
```
where:

url

Specifies the container registry URL with the docker:// prefix.

pullMethod

Optional: Set to node to use the node container runtime cache and pull secrets. The default value is pod.

secretRef

Optional: Specifies the name of a Secret object that stores the registry credentials.

certConfigMap

Optional: Specifies the name of a ConfigMap object that stores custom CA certificates for the registry.
Create the data volume by running the following command:
```
$ oc create -f <datavolume_manifest>.yaml
```
Monitor the import progress by running the following command:
```
$ oc get dv <datavolume_name>
```

Verification

Confirm that the data volume reaches the Succeeded phase by running the following command:

$ oc get dv <datavolume_name>

Example output:

NAME                        PHASE       PROGRESS
registry-image-datavolume   Succeeded   100.0%

Confirm that the PVC is created and bound by running the following command:

$ oc get pvc

Example output:

NAME                        STATUS   VOLUME     CAPACITY   ACCESS MODES
registry-image-datavolume   Bound    pv-name    20Gi       RWO

CPU architecture selection for `DataVolume` imports

You can target a specific CPU architecture when you import a virtual machine (VM) image from an Open Container Initiative (OCI) image index by setting the platform.architecture field in the DataVolume custom resource (CR).

warning

Specifying CPU architecture for data volume imports is a Technology Preview feature only. Technology Preview features are not supported with Red Hat production service level agreements (SLAs) and might not be functionally complete. Red Hat does not recommend using them in production. These features provide early access to upcoming product features, enabling customers to test functionality and provide feedback during the development process.

For more information about the support scope of Red Hat Technology Preview features, see Technology Preview Features Support Scope.

In heterogeneous clusters where nodes run different CPU architectures, importing the correct image variant is essential. Without specifying the target architecture, the imported image might not match the architecture of the node where the VM is scheduled to run. The platform.architecture field resolves this by selecting the appropriate variant from the image index.

The following CPU architectures are supported:

amd64
arm64
s390x

When you set the platform.architecture field in combination with pullMethod: node, OpenShift Virtualization adds a node selector to the importer pod. This node selector ensures that the importer pod runs on a node with the matching CPU architecture. The node pull method uses the container runtime cache of the node where the importer pod runs, so the pod must run on a node with the correct architecture. Without the node selector, the importer pod might run on a node with a different architecture, resulting in a mismatch between the pulled image and the target platform.

Example DataVolume CR:

apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: my-vm-image
spec:
  source:
    registry:
      url: "docker://registry.example.com/my-multi-arch-image:latest"
      pullMethod: node
      platform:
        architecture: arm64
  storage:
    resources:
      requests:
        storage: 30Gi

Import a VM image from an HTTP or HTTPS source

You can import a virtual machine (VM) image from an HTTP or HTTPS source by creating a DataVolume object. The Containerized Data Importer (CDI) controller creates a data volume and an underlying persistent volume claim (PVC), then downloads the image from the specified URL into the PVC.

Prerequisites

A VM image is accessible from an HTTP or HTTPS URL.
Optional: Custom CA certificates for HTTPS connections are stored in a ConfigMap object.
You have installed the OpenShift CLI (oc).

Procedure

Create a DataVolume object that specifies the HTTP or HTTPS source of the VM image:
```
apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: http-image-datavolume
spec:
  source:
    http:
      url: "https://example.com/images/my-vm-image.qcow2"
      secretRef: <http_secret>
      certConfigMap: <ca_certs_configmap>
  storage:
    resources:
      requests:
        storage: 20Gi
```
where:

url

Specifies the HTTP or HTTPS URL of the VM image.

secretRef

Optional: Specifies the name of a Secret object that stores Basic authentication credentials. The Secret must contain username and password keys.

certConfigMap

Optional: Specifies the name of a ConfigMap object that stores custom CA certificates for HTTPS connections.
Create the data volume by running the following command:
```
$ oc create -f <datavolume_manifest>.yaml
```
Monitor the import progress by running the following command:
```
$ oc get dv <datavolume_name>
```

Verification

Confirm that the data volume reaches the Succeeded phase by running the following command:

$ oc get dv <datavolume_name>

Example output:

NAME                     PHASE       PROGRESS
http-image-datavolume    Succeeded   100.0%

Confirm that the PVC is created and bound by running the following command:

$ oc get pvc

Example output:

NAME                     STATUS   VOLUME     CAPACITY   ACCESS MODES
http-image-datavolume    Bound    pv-name    20Gi       RWO

Import a VM image from S3 object storage

You can import a virtual machine (VM) image from an S3-compatible object storage bucket by using a DataVolume object. The Containerized Data Importer (CDI) pulls the image from the S3 endpoint and populates a persistent volume claim (PVC) that you can use to create a VM.

Prerequisites

A VM image is stored in an S3-compatible bucket.
You have S3 access credentials (access key ID and secret key) for your S3-compatible storage account.
You have installed the OpenShift CLI (oc).

Procedure

Create a Secret object for your S3 access credentials:
```
apiVersion: v1
kind: Secret
metadata:
  name: s3-secret
type: Opaque
stringData:
  accessKeyId: "<access_key>"
  secretKey: "<secret_key>"
```
where:

<access_key>

Specifies the access key ID for your S3-compatible storage account.

<secret_key>

Specifies the secret key for your S3-compatible storage account.
Apply the Secret object:
```
$ oc create -f <secret_manifest>.yaml
```
Create a DataVolume object that specifies the S3 source URL and secret reference:
```
apiVersion: cdi.kubevirt.io/v1beta1
kind: DataVolume
metadata:
  name: s3-image-datavolume
spec:
  source:
    s3:
      url: "https://s3.example.com/my-bucket/my-vm-image.qcow2"
      secretRef: s3-secret
      certConfigMap: <ca_certs_configmap>
  storage:
    resources:
      requests:
        storage: 20Gi
```
where:

url

Specifies the S3 URL of the VM image. For non-AWS S3 services, specify the full endpoint URL.

secretRef

Specifies the name of the Secret with the S3 access credentials.

certConfigMap

Optional: Specifies the name of a ConfigMap object that stores custom CA certificates for the S3 endpoint. Required only if your endpoint uses a custom certificate authority.
Apply the DataVolume object by running the following command:
```
$ oc create -f <datavolume_manifest>.yaml
```
Monitor the import progress by running the following command:
```
$ oc get dv <datavolume_name>
```

Verification

Confirm that the data volume reaches the Succeeded phase by running the following command:

$ oc get dv <datavolume_name>

Example output:

NAME                   PHASE       PROGRESS
s3-image-datavolume    Succeeded   100.0%

Confirm that the PVC is created and bound by running the following command:

$ oc get pvc

Example output:

NAME                   STATUS   VOLUME     CAPACITY   ACCESS MODES
s3-image-datavolume    Bound    pv-name    20Gi       RWO

`DataVolume` source fields

You can configure a DataVolume object to import a virtual machine (VM) image from several source types. Each source type has its own set of fields that control where and how the Containerized Data Importer (CDI) retrieves the image.

Registry source fields

The source.registry source type imports a VM image from a container registry.

source.registry** fields**

Field	Type	Description
`url`	String	Container registry URL of the image, specified with the `docker://` prefix. For example, `docker://registry.example.com/my-image:latest`.
`pullMethod`	String	Method for pulling the image. Specify `pod` to use the CDI importer pod or `node` to use the node’s container image cache. The default value is `pod`.
`secretRef`	String	Name of a `Secret` object that stores the credentials for the container registry.
`certConfigMap`	String	Name of a `ConfigMap` object that stores custom certificate authority (CA) certificates for the container registry.
`platform.architecture`	String	Target CPU architecture when importing from a multi-architecture image index. Supported values are `amd64`, `arm64`, and `s390x`. This field is a Technology Preview feature.

HTTP source fields

The source.http source type imports a VM image from an HTTP or HTTPS URL.

source.http** fields**

Field	Type	Description
`url`	String	HTTP or HTTPS URL of the VM image.
`secretRef`	String	Name of a `Secret` object that stores Basic authentication credentials for the HTTP server.
`certConfigMap`	String	Name of a `ConfigMap` object that stores custom certificate authority (CA) certificates for the HTTP server.
`extraHeaders`	List of strings	Additional HTTP headers to include in the image download request.
`secretExtraHeaders`	List of strings	Names of `Secret` objects that store additional HTTP headers to include in the image download request.

S3 source fields

The source.s3 source type imports a VM image from an S3-compatible object storage bucket.

source.s3** fields**

Field	Type	Description
`url`	String	S3 URL of the VM image.
`secretRef`	String	Name of a `Secret` object that stores the S3 access key ID and secret access key.
`certConfigMap`	String	Name of a `ConfigMap` object that stores custom certificate authority (CA) certificates for the S3 endpoint.

ImageIO source fields

The source.imageio source type imports a disk image from a Red Hat Virtualization (oVirt) environment by using the ImageIO API.

source.imageio** fields**

Field	Type	Description
`url`	String	URL of the oVirt Engine API endpoint. For example, `https://ovirt-engine.example.com/ovirt-engine/api`.
`diskId`	String	ID of the oVirt disk to import.
`secretRef`	String	Name of a `Secret` object that stores the credentials to access the oVirt Engine.
`certConfigMap`	String	Name of a `ConfigMap` object that stores CA certificates for the oVirt Engine.

VDDK source fields

The source.vddk source type imports a disk image from a VMware vSphere environment by using the Virtual Disk Development Kit (VDDK).

source.vddk** fields**

Field	Type	Description
`url`	String	URL of the vCenter or ESXi host.
`uuid`	String	UUID of the virtual machine in vCenter or ESXi.
`backingFile`	String	Path to the virtual hard drive to import from vCenter or ESXi.
`thumbprint`	String	TLS certificate thumbprint of the vCenter or ESXi host.
`secretRef`	String	Name of a `Secret` object that stores the username and password for the vCenter or ESXi host.
`initImageURL`	String	URL of a container image with an extracted VDDK library. Overrides the `v2v-vmware` config map.

Blank source

The source.blank source type creates an empty disk with no data. This source type has no configurable fields.

PVC source fields

The source.pvc source type clones the contents of an existing persistent volume claim (PVC).

source.pvc** fields**

Field	Type	Description
`namespace`	String	Namespace of the source PVC.
`name`	String	Name of the source PVC.

`VolumeSnapshot` source fields

The source.snapshot source type creates a volume from an existing VolumeSnapshot object.

source.snapshot** fields**

Field	Type	Description
`namespace`	String	Namespace of the source `VolumeSnapshot` object.
`name`	String	Name of the source `VolumeSnapshot` object.

Supported content types

The Containerized Data Importer (CDI) supports the following content types for VM image imports:

QCOW2
Raw
ISO
Archive (tar.gz)

Troubleshoot data volume imports

You can diagnose and resolve common issues with data volume imports by inspecting data volume conditions and events. Understanding these conditions helps you identify the root cause of failed or stalled imports.

Data volume conditions

Each data volume reports three conditions that indicate its current state. You can view these conditions by running the oc describe dv command.

Bound: Indicates whether the persistent volume claim (PVC) has been created and bound to a persistent volume. Status: True means the PVC is bound.
Running: Indicates whether an import operation is actively running. Status: True means data is being imported.
Ready: Indicates whether the data volume is ready to use. Status: True means the import completed successfully.

Diagnostic commands

You can use the following commands to gather information about a data volume import:

oc describe dv <datavolume_name>: Shows the data volume conditions, events, and status. Replace <datavolume_name> with the name of your data volume.
oc get events --field-selector involvedObject.name=<datavolume_name>: Shows events related to the data volume. Replace <datavolume_name> with the name of your data volume.

Common failure scenarios

Common data volume import failures

Symptom	Possible cause	Resolution
Import pod in `ErrImagePull` or `ImagePullBackOff`	Wrong registry URL or missing credentials	Verify the registry URL and ensure the `secretRef` points to a valid `Secret`.
Import pod in `CrashLoopBackOff`	Invalid image format or corrupted image	Verify the image is in a supported format (QCOW2, raw, ISO, archive).
Data volume stuck in `ImportScheduled`	No available persistent volume matching the storage request	Verify available storage and storage class configuration.
Data volume shows `Scratch space required` error	CDI requires scratch space for image conversion but none is available	Configure a scratch space storage class in the `CDI` object or ensure enough storage is available.
TLS certificate errors	Custom CA not configured	Create a `ConfigMap` object with the CA certificate and reference it in the `certConfigMap` field.
HTTP 401 or 403 errors	Missing or invalid authentication credentials	Verify the `Secret` referenced by `secretRef` has valid credentials.

Additional resources

Preparing CDI scratch space

Supported import sources​

Import a VM image from a container registry​

CPU architecture selection for DataVolume imports​

Import a VM image from an HTTP or HTTPS source​

Import a VM image from S3 object storage​

DataVolume source fields​

Registry source fields​

HTTP source fields​

S3 source fields​

ImageIO source fields​

VDDK source fields​

Blank source​

PVC source fields​

VolumeSnapshot source fields​

Supported content types​

Troubleshoot data volume imports​

Data volume conditions​

Diagnostic commands​

Common failure scenarios​

Supported import sources

Import a VM image from a container registry

CPU architecture selection for `DataVolume` imports

Import a VM image from an HTTP or HTTPS source

Import a VM image from S3 object storage

`DataVolume` source fields

Registry source fields

HTTP source fields

S3 source fields

ImageIO source fields

VDDK source fields

Blank source

PVC source fields

`VolumeSnapshot` source fields

Supported content types

Troubleshoot data volume imports

Data volume conditions

Diagnostic commands

Common failure scenarios