fix(nvlink): refresh NVL domain UUIDs for machines

[jayzhudev]

NvlPartitionMonitor refreshes NVL domain UUIDs for switches after each valid NMX-C Hello, but machines' NVL domain UUIDs were only populated when missing or nil. After an NVL domain UUID change, machines could therefore retain stale UUIDs.

This change updates machines' NVL domain UUIDs whenever a valid NMX-C Hello reports a different value.

Related issues

Supports #4397

Type of Change

• Add - New feature or capability
• Change - Changes in existing functionality
• Fix - Bug fixes
• Remove - Removed features or deprecated functionality
• Internal - Internal changes (refactoring, tests, docs, etc.)

Breaking Changes

• This PR contains breaking changes

Testing

• Unit tests added/updated
• Integration tests added/updated
• Manual testing performed
• No testing required (docs, internal refactor, etc.)

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 20e00854-f845-43d2-8859-ac64d38038cd

📥 Commits

Reviewing files that changed from the base of the PR and between 39d24a8 and ec35089.

📒 Files selected for processing (1)

• crates/nvlink-manager/src/lib.rs

🚧 Files skipped from review as they are similar to previous changes (1)

• crates/nvlink-manager/src/lib.rs

---

Summary by CodeRabbit

• Bug Fixes
  • Invalid nil domain UUIDs are now rejected during NMX-C Hello message processing.
  • Machine NVLink information now updates when the observed domain UUID changes.
  • Unchanged domain information is preserved without unnecessary updates.
  • Chassis serial numbers can now be recovered from previously saved machine data when unavailable in current snapshots.
  • Machines without a serial number in either source are safely skipped.

Walkthrough

The NVLink manager rejects nil Hello domain UUIDs. It updates machine NVLink information when the domain changes or chassis serial data is missing. It falls back to persisted serial data and skips machines without serial data. Tests cover these cases.

Changes

NVLink domain and serial synchronization

Layer / File(s)	Summary
Hello UUID validation and assignment crates/nvlink-manager/src/lib.rs	Hello parsing rejects malformed and nil domain UUIDs. Existing machine NVLink information receives the validated domain UUID.
Machine NVLink refresh and serial fallback crates/nvlink-manager/src/lib.rs	Refresh logic updates changed domain data, resolves chassis serials from snapshots or persisted data, skips machines without either source, and preserves unchanged data. Tests cover the supported cases.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Possibly related PRs

• NVIDIA/infra-controller#4476: Both changes modify NVLink domain UUID handling and propagation.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main change: refreshing NVLink domain UUIDs for machines.
Description check	✅ Passed	The description directly explains the stale UUID bug, the fix, related issue, and testing performed.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@crates/nvlink-manager/src/lib.rs`:
- Line 236: Update process_nmx_c_partition_monitor_group and the
existing-machine comparison around MachineNvLinkInfo so a missing snapshot
chassis serial falls back to the persisted chassis serial, allowing stale domain
UUIDs to be detected and refreshed. Add a scenario covering an absent snapshot
serial with a differing persisted domain UUID, and verify the machine
information is updated.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: b8dcb6e5-8e94-43ad-a4b6-700d8f85f500

📥 Commits

Reviewing files that changed from the base of the PR and between 92c363a and 39d24a8.

📒 Files selected for processing (1)

• crates/nvlink-manager/src/lib.rs