docs(boot-interface): document how a queued NIC-mode flip applies#3648
docs(boot-interface): document how a queued NIC-mode flip applies#3648chet wants to merge 1 commit into
Conversation
Flipping a DPU to NIC mode has been fully automatic for a while -- site-explorer power-cycles the host on every vendor, escalating a refused `PowerCycle` to a cold `ACPowercycle` -- but the operator guide never said so, and the one failure state had no troubleshooting entry. Section 3.5 now walks the application mechanics (the staged `Mode.Set`, the automatic power cycle, the rate limit), and the Section 8 table gains a row for the `manual_power_cycle_required` pairing blocker with the manual recovery step. This lands against the current guide; the in-flight refresh in NVIDIA#3270 touches the same file, so whichever merges second rebases -- this delta is two additions and slots cleanly either way. This supports NVIDIA#3644 Signed-off-by: Chet Nichols III <chetn@nvidia.com>
|
@coderabbitai full_review, thanks! |
|
✅ Action performedFull review finished. |
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (1)
Summary by CodeRabbit
WalkthroughThe operator guide documents queued DPU mode changes, site-explorer power-cycle escalation, rate-limited exploration, and the ChangesDPU mode flip documentation
Estimated code review effort: 1 (Trivial) | ~2 minutes Possibly related issues
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
|
🌿 Preview your docs: https://nvidia-preview-pull-request-3648.docs.buildwithfern.com/infra-controller |
🔍 Container Scan SummaryNo Grype artifacts were found to aggregate. |
There was a problem hiding this comment.
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 `@docs/provisioning/boot-interfaces-and-dpu-modes.md`:
- Line 341: Revise the manual_power_cycle_required guidance to prioritize
waiting for the next verification/convergence window while the host may still be
mid-flip or within reset_rate_limit. Instruct operators to manually power-cycle
only if the host remains in the old mode after that window or reset attempts
continue failing, and retain the existing recovery commands and blocker-clearing
behavior.
🪄 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: 9fdf3930-7b9b-4987-9389-a4e538b2ff16
📒 Files selected for processing (1)
docs/provisioning/boot-interfaces-and-dpu-modes.md
Section 3.5 of the boot-interface guide now documents how a queued NIC-mode flip actually applies -- the staged
Mode.Set, site-explorer's automatic power cycle on every vendor (PowerCycle, escalating to coldACPowercyclewhere refused), and the rate limit -- and the Section 8 troubleshooting table gains a row for themanual_power_cycle_requiredpairing blocker with the manual recovery step.Sequenced with #3270 (same file): whichever merges second rebases.
This supports #3644