Is this page helpful?

VM Application Deployment#

OpenShift Virtualization extends Red Hat OpenShift with the KubeVirt APIs and controllers required to run virtual machines (VMs) with container workloads on the same cluster. Kubernetes resources represent a VM, and you manage it through the OpenShift web console, the oc command-line tool, or virtctl.

The VM operating-system disk resides in a PersistentVolumeClaim (PVC). For this workflow, the source disk is a QCOW2 image. QCOW2 is a sparse, copy-on-write disk format that supports efficient local image development and snapshots. The operating-system-specific procedure uploads the image to a PVC. The common deployment procedure creates a DataSource that references the uploaded PVC, and OpenShift Virtualization clones the source disk when you create each VM.

The general workflow is:

  1. Prepare a Windows or Linux QCOW2 image.

  2. Install the operating-system drivers and application software in the image.

  3. Shut down the builder VM and export a versioned QCOW2 image.

  4. Upload the image to a versioned PVC.

  5. Create a DataSource that references the uploaded PVC.

  6. Create Windows or Linux VMs from the DataSource and attach the required GPU, network, and storage resources.

For more information about the underlying virtualization platform, refer to the OpenShift Virtualization documentation.