Skip to content

ci(os-49): add temporary VM validation workflow#1174

Closed
jtoelke2 wants to merge 4 commits intomainfrom
os-49-vm-validation/jt
Closed

ci(os-49): add temporary VM validation workflow#1174
jtoelke2 wants to merge 4 commits intomainfrom
os-49-vm-validation/jt

Conversation

@jtoelke2
Copy link
Copy Markdown
Collaborator

@jtoelke2 jtoelke2 commented May 5, 2026

Summary

Temporary OS-49 validation PR. This adds a throwaway workflow to collect evidence for VM boot support and non-destructive GPU/VFIO host capability on shared runners. It is not intended to merge.

Related Issue

OS-49 (Linear)

Changes

  • Adds OS-49 VM Validation workflow triggered by pull-request/<N> mirror pushes or manual dispatch.
  • Runs mise run e2e:vm on linux-amd64-cpu8 and linux-arm64-cpu8 to test actual VM boot on shared CPU runners.
  • Probes shared Linux GPU runners for /dev/kvm, NVIDIA PCI devices, IOMMU groups, and VFIO sysfs readiness without binding or unbinding GPUs.

Testing

  • Workflow YAML parses locally with Ruby YAML loader.
  • mise run pre-commit passes. It was run, but current main fails ssh::tests::launch_editor_returns_friendly_error_when_binary_missing; rerunning that single test reproduced the same unrelated failure.
  • E2E tests added/updated if applicable. This PR adds a temporary CI validation workflow; evidence comes from the PR workflow run.

Checklist

  • Follows Conventional Commits
  • Commit is signed off (DCO)
  • Architecture docs updated (not applicable for temporary validation PR)

Cleanup

Close this PR after collecting the VM boot and GPU/VFIO probe results. Do not merge unless the team explicitly decides to keep a permanent version of this workflow.

Signed-off-by: Jonas Toelke <jtoelke@nvidia.com>
@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot Bot commented May 5, 2026

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

@jtoelke2 jtoelke2 marked this pull request as ready for review May 5, 2026 18:06
jtoelke2 added 3 commits May 5, 2026 13:14
Signed-off-by: Jonas Toelke <jtoelke@nvidia.com>
Signed-off-by: Jonas Toelke <jtoelke@nvidia.com>
Signed-off-by: Jonas Toelke <jtoelke@nvidia.com>
@jtoelke2
Copy link
Copy Markdown
Collaborator Author

jtoelke2 commented May 6, 2026

Closing this evidence-only PR now that the OS-49 VM runner evidence is captured in Linear. It was never intended to merge: shared CPU labels do not expose /dev/kvm, shared GPU can run non-passthrough e2e:vm, and probed shared GPU labels do not expose writable vfio-pci controls for passthrough.

@jtoelke2 jtoelke2 closed this May 6, 2026
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.

1 participant