Skip to content

Update the building of vgpu manager build doc #286

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Oct 18, 2025
Merged

Update the building of vgpu manager build doc #286

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
69d4b97
Update the building of vgpu manager build doc
Oct 16, 2025
fb17003
Minor changes for style guide compliance
chenopis Oct 17, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
31 changes: 12 additions & 19 deletions gpu-operator/gpu-operator-kubevirt.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -537,43 +537,36 @@ Open a terminal and clone the driver container image repository.
$ git clone https://github.com/NVIDIA/gpu-driver-container.git
$ cd gpu-driver-container

Change to the vgpu-manager directory for your OS. We use Ubuntu 20.04 as an example.
#. Copy the NVIDIA vGPU manager from your extracted ZIP file to the operating system version you want to build the image for:
* We use Ubuntu 22.04 as an example.

.. code-block:: console

$ cd vgpu-manager/ubuntu20.04

.. note::
Copy ``<local-driver-download-directory>/\*-vgpu-kvm.run`` to ``vgpu-manager/ubuntu22.04/``.

For Red Hat OpenShift, run ``cd vgpu-manager/rhel8`` to use the ``rhel8`` folder instead.
.. code-block:: console

Copy the NVIDIA vGPU Manager from your extracted zip file
$ cp <local-driver-download-directory>/*-vgpu-kvm.run vgpu-manager/ubuntu22.04/

.. code-block:: console
.. note::

$ cp <local-driver-download-directory>/*-vgpu-kvm.run ./
For Red Hat OpenShift, use a directory that includes ``rhel`` in the directory name. For example, ``vgpu-manager/rhel8``.

| Set the following environment variables:
| ``PRIVATE_REGISTRY`` - name of private registry used to store driver image
| ``VERSION`` - NVIDIA vGPU Manager version downloaded from NVIDIA Software Portal
| ``OS_TAG`` - this must match the Guest OS version. In the following example ``ubuntu20.04`` is used. For Red Hat OpenShift this should be set to ``rhcos4.x`` where x is the supported minor OCP version.
| ``CUDA_VERSION`` - CUDA base image version to build the driver image with.
| ``VGPU_HOST_DRIVER_VERSION`` - NVIDIA vGPU Manager version downloaded from NVIDIA Software Portal
| ``OS_TAG`` - this must match the Guest OS version. In the following example ``ubuntu22.04`` is used. For Red Hat OpenShift this should be set to ``rhcos4.x`` where x is the supported minor OCP version.

.. code-block:: console

$ export PRIVATE_REGISTRY=my/private/registry VERSION=510.73.06 OS_TAG=ubuntu20.04 CUDA_VERSION=11.7.1
$ export PRIVATE_REGISTRY=my/private/registry VGPU_HOST_DRIVER_VERSION=580.82.07 OS_TAG=ubuntu22.04

Build the NVIDIA vGPU Manager image.

.. code-block:: console

$ docker build \
--build-arg DRIVER_VERSION=${VERSION} \
--build-arg CUDA_VERSION=${CUDA_VERSION} \
-t ${PRIVATE_REGISTRY}/vgpu-manager:${VERSION}-${OS_TAG} .
$ VGPU_HOST_DRIVER_VERSION=${VGPU_HOST_DRIVER_VERSION} IMAGE_NAME=${PRIVATE_REGISTRY}/vgpu-manager make build-vgpuhost-${OS_TAG}

Push NVIDIA vGPU Manager image to your private registry.

.. code-block:: console

$ docker push ${PRIVATE_REGISTRY}/vgpu-manager:${VERSION}-${OS_TAG}
$ VGPU_HOST_DRIVER_VERSION=${VGPU_HOST_DRIVER_VERSION} IMAGE_NAME=${PRIVATE_REGISTRY}/vgpu-manager make push-vgpuhost-${OS_TAG}
27 changes: 11 additions & 16 deletions openshift/openshift-virtualization.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -248,28 +248,25 @@ Use the following steps to build the vGPU Manager container and push it to a pri
$ git clone https://github.com/NVIDIA/gpu-driver-container.git
$ cd gpu-driver-container

#. Change to the ``vgpu-manager`` directory for your OS:
#. Copy the NVIDIA vGPU manager from your extracted ZIP file to the operating system version you want to build the image for:
* We use RHEL 8 as an example.

.. code-block:: console

$ cd vgpu-manager/rhel8

#. Copy the NVIDIA vGPU Manager from your extracted zip file:
Copy ``<local-driver-download-directory>/\*-vgpu-kvm.run`` to ``vgpu-manager/rhel8/``.

.. code-block:: console

$ cp <local-driver-download-directory>/*-vgpu-kvm.run ./
$ cp <local-driver-download-directory>/*-vgpu-kvm.run vgpu-manager/rhel8/

#. Set the following environment variables.

* ``PRIVATE_REGISTRY`` - Name of the private registry used to store the driver image.
* ``VERSION`` - The NVIDIA vGPU Manager version downloaded from the NVIDIA Software Portal.
* ``VGPU_HOST_DRIVER_VERSION`` - The NVIDIA vGPU Manager version downloaded from the NVIDIA Software Portal.
* ``OS_TAG`` - This must match the Guest OS version.
For RedHat OpenShift, specify ``rhcos4.x`` where _x_ is the supported minor OCP version.

.. code-block:: console

$ export PRIVATE_REGISTRY=my/private/registry VERSION=510.73.06 OS_TAG=rhcos4.11
$ export PRIVATE_REGISTRY=my/private/registry VGPU_HOST_DRIVER_VERSION=580.82.07 OS_TAG=rhcos4.18

.. note::

Expand All @@ -280,15 +277,13 @@ Use the following steps to build the vGPU Manager container and push it to a pri

.. code-block:: console

$ docker build \
--build-arg DRIVER_VERSION=${VERSION} \
-t ${PRIVATE_REGISTRY}/vgpu-manager:${VERSION}-${OS_TAG} .
$ VGPU_HOST_DRIVER_VERSION=${VGPU_HOST_DRIVER_VERSION} IMAGE_NAME=${PRIVATE_REGISTRY}/vgpu-manager make build-vgpuhost-${OS_TAG}

#. Push the NVIDIA vGPU Manager image to your private registry.

.. code-block:: console

$ docker push ${PRIVATE_REGISTRY}/vgpu-manager:${VERSION}-${OS_TAG}
$ VGPU_HOST_DRIVER_VERSION=${VGPU_HOST_DRIVER_VERSION} IMAGE_NAME=${PRIVATE_REGISTRY}/vgpu-manager make push-vgpuhost-${OS_TAG}

.. _install-the-gpu-operator:

Expand Down Expand Up @@ -425,7 +420,7 @@ As a cluster administrator, you can create a ClusterPolicy using the OpenShift C
* Under *env*, fill in **image** with ``vgpu-manager`` and the **version** with your driver version.
* Expand the **NVIDIA vGPU Device Manager config** section and make sure that the **enabled** checkbox is checked.

If you are only using GPU passthrough, you don't need to fill these sections out.
If you are only using GPU passthrough, you do not need to fill these sections out.

* Expand the **VFIO Manager config** section and select the **enabled** checkbox.
* Optionally, in the **Sandbox Workloads config** section, set **defaultWorkload** to ``vm-passthrough`` if you want passthrough to be the default mode.
Expand Down Expand Up @@ -687,7 +682,7 @@ Switching vGPU device configuration after one has been successfully applied assu

To apply a new configuration after GPU Operator install, simply update the ``nvidia.com/vgpu.config`` node label.

Let's run through an example on a system with two **A10** GPUs.
The following example shows a system with two **A10** GPUs.

.. code-block:: console

Expand All @@ -704,7 +699,7 @@ After installing the GPU Operator as detailed in the previous sections and witho
"nvidia.com/NVIDIA_A10-12Q": "4"
}

If instead you want to create **A10-4Q** devices, we can label the node like such:
If instead you want to create **A10-4Q** devices, label the node as follows:

.. code-block:: console

Expand Down