docs: use CUDA base image for GPU smoke test in spark-install

[latenighthackathon]

Summary

• Replace ubuntu with nvidia/cuda:12.8.0-base-ubuntu24.04 in the GPU verification command

Related Issue

Closes #1166

Changes

The GPU smoke test in spark-install.md uses the ubuntu Docker image, which does not include nvidia-smi. The command always fails even when GPUs are correctly configured, making it useless as a validation step.

Switched to nvidia/cuda:12.8.0-base-ubuntu24.04 which includes nvidia-smi and correctly validates GPU access through the NVIDIA container runtime.

Testing

• Verified nvidia/cuda:12.8.0-base-ubuntu24.04 exists on Docker Hub
• Standard NVIDIA GPU validation image used in NVIDIA documentation

Checklist

• Conventional commit format
• Scoped to issue, no unrelated changes
• No secrets or credentials

Signed-off-by: latenighthackathon latenighthackathon@users.noreply.github.com

Summary by CodeRabbit

• Documentation
  • Clarified NVIDIA container runtime verification instructions to specify an explicit CUDA base image version for clearer, more consistent GPU-enabled container setup. The verification step and GPU check remain intact, improving reproducibility of the documented workflow.

[coderabbitai]

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

• @coderabbitai resume to resume automatic reviews.
• @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

• ▶️ Resume reviews
• 🔍 Trigger review

📝 Walkthrough

Walkthrough

The GPU smoke-test Docker command in documentation now uses the nvidia/cuda:12.8.0-base-ubuntu24.04 image instead of ubuntu, so nvidia-smi is executed inside a container that includes the binary.

Changes

Cohort / File(s)	Summary
Documentation Fix spark-install.md	Replaced ubuntu with nvidia/cuda:12.8.0-base-ubuntu24.04 in the GPU smoke-test docker run --rm --runtime=nvidia --gpus all ... nvidia-smi command so the nvidia-smi binary exists inside the container.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~2 minutes

Poem

🐰
I hopped in docs with nimble paw,
Swapped base to CUDA—fixed the flaw.
Now nvidia-smi wakes with glee,
The smoke test hums—no mystery! 🥕

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main change: replacing the ubuntu image with a CUDA base image for the GPU smoke test in spark-install documentation.
Linked Issues check	✅ Passed	The PR directly addresses the primary objective from issue #1166 by replacing the ubuntu image with nvidia/cuda:12.8.0-base-ubuntu24.04, which includes nvidia-smi and enables proper GPU validation.
Out of Scope Changes check	✅ Passed	All changes are scoped to the GPU smoke test command in spark-install.md and directly address the documented issue; no out-of-scope modifications detected.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[wscurran]

✨ Thanks for submitting this pull request, which proposes a way to improve the documentation for the GPU smoke test in spark-install. This could help users verify their GPU configuration more accurately.

---

Possibly related open issues:

• #1166 [NemoClaw][spark 1.120.38] [doc] spark-install .md command fails because ubuntu image lacks nvidia-smi
• #1117 sandbox@my-assistant:~$ openclaw update ## npm install -g openclaw@latest
• #1260 Multi-agent deployments need workspace sync strategy — docs assume single agent

[prekshivyas]

@latenighthackathon thanks ! pls update your branch, add a sign-off in your PR description, otherwise this looks good for approval

[prekshivyas]

@coderabbitai review

[coderabbitai]

✅ Actions performed

Review triggered.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[latenighthackathon]

@prekshivyas @wscurran Thanks for the review! I've rebased onto latest origin/main — single clean commit now (f86b936a). The Signed-off-by trailer is on both the commit and at the bottom of the PR body. Let me know if you'd like it somewhere else in the description.

Cheers!

[prekshivyas]

Correct — ubuntu image doesn't have nvidia-smi. CUDA base image is the standard for GPU validation. @latenighthackathon can you rebase onto main?

[latenighthackathon]

@prekshivyas Rebased onto latest origin/main — single clean commit (521dff6f), merge commit dropped. Sign-off present on both commit and PR body.

Cheers!

[prekshivyas]

@latenighthackathon is it okay if you can give us access to update branches on your fork ?
would help speed up the merge requests you have - most of them are ready for approval but blocked on Update Branch

[latenighthackathon]

@prekshivyas Done — flipped "Allow edits by maintainers" on my open NemoClaw PRs (#1226, #1725, #1286, #1897, #1898) and on OpenShell #871 while I was at it. You should be able to Update Branch directly on any of them now.

Cheers!