-
Notifications
You must be signed in to change notification settings - Fork 225
GPU: Add unsupported distro list, add NVIDIA GPU driver installation for Debian #4132
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.
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
Conversation
b35f262 to
cc89971
Compare
|
Test is ongoing. The PR potentially impacts a lot of distro and variants. |
ecf55d7 to
447f69c
Compare
447f69c to
6c67f3f
Compare
|
Not include changes for SUSE in this PR due to multiple issues faced while enabling SUSE. It may need more time |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull request overview
This PR adds NVIDIA GPU driver installation support for Debian distributions and improves the overall GPU driver installation architecture. It refactors the Ubuntu installation to share common installation logic with Debian through a new _install_cuda_debian_based helper method, and improves test coverage by explicitly declaring unsupported operating systems in test requirements rather than checking at runtime.
Key Changes
- Added Debian support for NVIDIA CUDA driver installation with version checks for Debian 10, 11, and 12
- Refactored Ubuntu and Debian installation methods to share common logic via
_install_cuda_debian_based - Changed
_install_package_from_urlfrom private to public method (install_package_from_url) in the Posix operating system class - Moved GPU test OS compatibility checks from runtime validation to declarative test requirements
Reviewed changes
Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
| lisa/tools/gpu_drivers.py | Added Debian OS support, new constants for CUDA repos, refactored Ubuntu/Debian installation to share common logic, updated dependency installation for Debian, and changed private method calls to public |
| lisa/operating_system.py | Changed _install_package_from_url method visibility from private to public |
| lisa/microsoft/testsuites/gpu/gpusuite.py | Added explicit unsupported OS lists (AlmaLinux, Oracle, Suse) to test requirements and removed redundant runtime OS checks from _check_driver_installed |
| lisa/microsoft/testsuites/cvm/cvm_attestation_tool.py | Updated method call from private _install_package_from_url to public install_package_from_url |
| lisa/sut_orchestrator/azure/features.py | Updated TODO comment to clarify separation of concerns between cloud and OS-level GPU requirements |
Testing Recommendations (LISA Guidelines):
Key Test Cases:
verify_load_gpu_driver|verify_gpu_extension_installation|verify_gpu_cuda_with_pytorch
Impacted LISA Features:
Gpu
Tested Azure Marketplace Images:
debian debian-11 11-gen2 latest
debian debian-12 12 latest
debian debian-12 12-gen2 latest
canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest
canonical ubuntu-24_04-lts server latest
|
@adityagesh please rebase the code, also let know once the PR is ready for review. |
|
Test fail when run against |
Move logic to skip GPU cases based on distro to case requirement
Checking of GPU support is not supposed to happen in verification process. The is_supported checks are spread across multiple functions for GPU. We need to gradually reduce this incosistency
6c67f3f to
85e67fb
Compare
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
While raising the PR, the cuda packages weren't available. Currently the package is available, but post installation the driver doesn't seem to be correctly installed. I think the package for Debian 13 is still not fully ready. |
No description provided.