Repository navigation
epic(cluster): firmware / BIOS / BMC audit & standardization #1081
Description
Firmware / bios / bmc audit & standardization
Purpose
Every core machine on identical, known firmware and settings, verifiable on a routine basis without walking to the basement.
Scope
Audit BIOS and BMC/IPMI ('ASM') versions and every setting on the core machines; bring them into line (WRX80 Sage already updated; Dell machines had no firmware updates found - best effort recorded). Build the repeatable check: a system that knows every KVM IP, screenshots and verifies settings routinely, and emits a reappliable config. Output lands in developing-today/hardware-doc as VERIFICATION records.
Execution mode
Hybrid: version/setting comparison and tooling are online; machines must be powered and on the management network (physical), and some settings can only be changed at the console.
Dependencies
Depends on epic 1 (inventory of machines and KVM addresses).
Acceptance
Per-machine firmware/setting records committed; a diff showing core machines identical or every exception named; the routine screenshot-verification check runs and its output is filed.
Structure
Part of the basement cluster program. Project home: planning/roadmaps/basement-cluster.md in this repo. Full program register: goal workspace files/idea-register-2026-10-20.md (Muse goal goal_ba17a5777c21). Detailed task issues are filed as sub-issues of this epic; dependencies use blocks / blocked-by links.
Hard rules for the whole program: no HashiCorp anything; everything bootstrappable from config, no hand-built state.
Activity
developing-today commentedon Oct 27, 2026
CODE-223
dezren39 commentedon Oct 27, 2026
Epic map — basement cluster program
This epic: #1081 — Firmware / BIOS / BMC audit
Dependencies: Blocked by #1080 (machine/KVM inventory)
Filed in the 2026-10-26 cluster seed. Project home:
planning/roadmaps/basement-cluster.md.docs(basement-cluster): link filed epic issues in roadmap
dezren39 commentedon Oct 27, 2026
Predecessor: this epic supersedes dezren39/darwin-nix#502 (filed 2026-10-20 in the nix repo during the program's first filing pass). A superseded-pointer comment is on that issue; work happens here.
dezren39 commentedon Nov 3, 2026
agent-note:v1
state: triaged (coverage pass)
Coverage pass 2026-11-03: reviewed in the fleet-wide per-issue pass (drift / overlap / already-handled / stale / blockers / ordering). Verdict: still live; queued in repo-health-code-pilot as triage at priority 30 (priority/none). If this is already handled at current main, the triage closes it with evidence; if it needs splitting, the triage files the split.
Found via: coverage pass 2026-11-03
dezren39 commentedon Nov 3, 2026
🤖 Agent note · triaged — still live, split already exists, dependency-blocked
Verdict: issue #1081 is still live at current main — not handled, not superseded (it is itself the successor of dezren39/darwin-nix#502). No fix path is executable yet; the blockers and ordering below are the triage outcome.
Structure (already filed, no new split needed):
Blocker / ordering: this epic depends on epic #1080 (hardware inventory), which is still OPEN: its schema task #1094 is closed, but #1095 (core Threadripper stats), #1096 (remaining machines), #1097 (NIC/switch facts) and #1098 (rank + first-node selection) are all OPEN. Task #1099 cannot collect per-machine firmware/settings records until the inventory names the machines and their KVM/management addresses. Execution is also mode/hybrid: machines must be powered and on the management network, and some settings are console-only — so even after the inventory lands, #1099/#1101 need physical/console access windows, not just an agent session.
Next action: finish #1095–#1097 (inventory data), then unblock task #1099; #1100 can be drafted in parallel from vendor docs but finalizes only after #1099's records exist.