Skip to content

Fix bug in NVIDIA Driver install script when package version is specified#782

Open
cdesiniotis wants to merge 2 commits intomainfrom
fix-driver-package-version
Open

Fix bug in NVIDIA Driver install script when package version is specified#782
cdesiniotis wants to merge 2 commits intomainfrom
fix-driver-package-version

Conversation

@cdesiniotis
Copy link
Copy Markdown

@cdesiniotis cdesiniotis commented Apr 21, 2026

Before this change, when nvidiaDriver.package.version was specified,
holodeck was not populating the package version correctly. As a result,
the package installation failed. For example, one would see error messages
like:

  sudo DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends cuda-drivers=580.95.05
  E: Version '580.95.05' for 'cuda-drivers' was not found

This commit ensures we construct the package version string correctly
by adding the '-*' suffix to the version string.

Fixes #783

@coveralls
Copy link
Copy Markdown

coveralls commented Apr 21, 2026

Coverage Report for CI Build 24748725397

Coverage remained the same at 47.77%

Details

  • Coverage remained the same as the base build.
  • Patch coverage: No coverable lines changed in this PR.
  • No coverage regressions found.

Uncovered Changes

No uncovered changes found.

Coverage Regressions

No coverage regressions found.


Coverage Stats

Coverage Status
Relevant Lines: 11032
Covered Lines: 5270
Line Coverage: 47.77%
Coverage Strength: 0.53 hits per line

💛 - Coveralls

Copy link
Copy Markdown
Collaborator

@ArangoGutierrez ArangoGutierrez left a comment

Choose a reason for hiding this comment

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

Thanks for the PR Chris! I truly appreciate it :)

Comment on lines +24 to +25
package:
version: 580.95.05
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

The current NVIDIADriver API on main (api/holodeck/v1alpha1/types.go:433) exposes driver version as a top-level field:

  nvidiaDriver:                                                                                                                                                                                                                            
    install: true                                                                                                                                                                                                                          
    version: 580.95.05

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

I made this change because of this comment which says the top-level field is deprecated in favor of nvidiaDriver.package.version:

// Version is deprecated, use Package.Version instead.

Comment thread tests/data/test_aws_driver_package.yml Outdated
version: 580.95.05
kubernetes:
install: true
installer: kubeadm No newline at end of file
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

Missing EOF

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

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

Updated!

@cdesiniotis cdesiniotis force-pushed the fix-driver-package-version branch from 33e02d8 to 84488da Compare April 21, 2026 20:56
@cdesiniotis cdesiniotis changed the title Add e2e test that specifies NVIDIA driver package version Fix bug in NVIDIA Driver install script when package version is specified Apr 21, 2026
@cdesiniotis cdesiniotis force-pushed the fix-driver-package-version branch 2 times, most recently from 6cbaa4d to 4b035b2 Compare April 21, 2026 21:18
@cdesiniotis cdesiniotis marked this pull request as ready for review April 21, 2026 21:28
@cdesiniotis cdesiniotis requested a review from tariq1890 April 21, 2026 21:28
Signed-off-by: Christopher Desiniotis <cdesiniotis@nvidia.com>
…fied

Before this change, when nvidiaDriver.package.version was specified,
holodeck was not populating the package version correctly. As a result,
the package installation failed. For example, one would see error messages
like:

  sudo DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends cuda-drivers=580.95.05
  E: Version '580.95.05' for 'cuda-drivers' was not found

This commit ensures we construct the package version string correctly
by adding the '-*' suffix to the version string.

Fixes #783

Signed-off-by: Christopher Desiniotis <cdesiniotis@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Bug]: nvidiaDriver.package.version is not functional

3 participants