Skip to content

feat(machine-a-tron): generate unique InfiniBand GUIDs for machine-a-tron hosts #3518

Description

@poroh

Is this a new feature, an enhancement, or a change to existing functionality?

New Feature

How would you describe the priority of this feature request

Medium

Please provide a clear description of problem this feature solves

Some bmc-mock hardware models return fixed InfiniBand GUIDs. H100 GUIDs repeat across hosts, while GB200 models can repeat GUIDs within a single host.

UFM and NICo identify ports by GUID. Duplicate values prevent simulated ports from being represented accurately and make discovery ambiguous when real and simulated fabrics coexist.

Feature Description

When machine-a-tron creates an IB-capable host, its simulated hardware shall report a deterministic GUID that is unique for every IB interface across all simulated hosts.

Describe your ideal solution

Derive each GUID from stable per-host data, such as the host serial or allocated hardware address range, plus the interface index.

Implement GUID generation in the relevant bmc-mock hardware models so Redfish discovery and machine-a-tron observe the same values. Preserve the expected 64-bit hexadecimal representation.

Describe any alternatives you have considered

No response

Additional context

No response

Code of Conduct

  • I agree to follow NVIDIA Infra Controller's Code of Conduct
  • I have searched the open feature requests and have found no duplicates for this feature request

Metadata

Metadata

Assignees

No one assigned

    Labels

    internalinternal work transparent to end-userscaling

    Fields

    No fields configured for Enhancement.

    Projects

    Status
    Triage

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions