Skip to content

Update GPU related docs #538

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
sureshanaparti merged 1 commit into from
Aug 14, 2025
Merged

Update GPU related docs #538

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
28 changes: 26 additions & 2 deletions source/adminguide/hosts.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -243,10 +243,34 @@ the host.

The cloudstack agent uses the ``gpudiscovery.sh`` script to discover the GPU
devices on a KVM host. The script is located in the
``/usr/share/cloudstack-common/scripts/vm/`` directory on the host.
``/usr/share/cloudstack-common/scripts/vm/`` directory on the host. The script
relies on the ``lspci`` & ``xmlstarlet`` command to discover the GPU devices
and their status on the host. So, for the discovery to be successful, the
``lspci`` & ``xmlstarlet`` commands needs to be installed on the host.
Copy link

Copilot AI Aug 14, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Subject-verb disagreement: 'commands needs' should be 'commands need' (plural subject requires plural verb).

Suggested change
``lspci`` & ``xmlstarlet`` commands needs to be installed on the host.
``lspci`` & ``xmlstarlet`` commands need to be installed on the host.

Copilot uses AI. Check for mistakes.

.. parsed-literal::

dnf install pciutils xmlstarlet

.. parsed-literal::

sudo apt install pciutils xmlstarlet

.. note::
The script can be run manually to debug the discovery of the GPU devices on a host.
The following table shows the compatibility matrix for NVIDIA vGPU types with CloudStack:

.. cssclass:: table-striped table-bordered table-hover

=============================== ================== =======================
NVIDIA vGPU Type VFIO Framework Supported in CloudStack
=============================== ================== =======================
Legacy: SR-IOV not supported `mdev` Yes
SR-IOV supported `mdev` Yes
SR-IOV supported `Vendor specific` No
Multi Instance GPU No
=============================== ================== =======================

The script can also be run manually to debug the discovery of the GPU devices on a host.

.. parsed-literal::

Expand Down
4 changes: 3 additions & 1 deletion source/adminguide/service_offerings.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -304,7 +304,9 @@ To create a new compute offering:
This is applicable only for KVM hypervisor.

- **GPU Display**: Whether to use the GPU device attached to the guest instance for display.
This is applicable only for KVM hypervisor.
This is applicable only for KVM hypervisor. Depending on the OS and display configuration,
the user might need to set ``video.hardware`` to ``none`` in the instance's settings to
use CPVM for display.
Copy link

Copilot AI Aug 14, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The documentation should clarify where exactly the instance settings are located and how to access them. Consider adding the specific path or UI location where users can modify the video.hardware setting.

Suggested change
use CPVM for display.
use CPVM for display. To modify the ``video.hardware`` setting, navigate to the instance's details page in the CloudStack UI, click on the "Settings" tab, and locate the ``video.hardware`` option to change its value.

Copilot uses AI. Check for mistakes.

- **Public**: Indicate whether the compute offering should be
available to all domains or only some domains. Choose Yes to make it
Expand Down