docs: SLICE 4 — hub README, REPORT, CONTEXT (offsite arc build-complete)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
This commit is contained in:
+13
@@ -3,6 +3,19 @@
|
||||
> Created with the REUSE.md rollout (2026-07-03). Authoritative history: `hub/CHANGELOG.md` (hub),
|
||||
> `website/CHANGELOG.md`, `scripts/CHANGELOG.md`; end-of-task detail in `REPORT.md`.
|
||||
|
||||
- **2026-07-10 — SLICE 4 soft-quota + offsite health SHIPPED+LIVE (hub v0.41.0 + controller v0.109.1) —
|
||||
the offsite arc is BUILD-COMPLETE.** Controller: `quota_gb` rides the descriptor into the target (hash
|
||||
includes it → quota changes re-apply via key-auth-first, no consume); `repo_size_bytes` from restic
|
||||
stats; pre-run gate (≥100% refuses NEW backups Hungarian-loudly, **prune still runs** — red-proofed;
|
||||
restore never gated; ≥80% warns); /backups usage bar. Hub: `monitor.OffsiteChecker` (fill 90/95 of
|
||||
quota + `offsite_stale` >48h for enabled+escrowed only — red-proofed; nil-safe on old reports) +
|
||||
**manual-only Freeze/Unfreeze lever** (readonly on the exactly-1 labelled sub-account, SSH preserved).
|
||||
**Live:** key-auth-first proven against the real box (the deploy's own re-apply); that re-apply exposed
|
||||
+ fixed v0.109.1 (re-apply must preserve EscrowState + runtime status — it demoted the escrowed demo);
|
||||
demo repaired (manual confirm + run: escrowed, 45 MB / 50 GB, checker healthy 0 events, usage bar
|
||||
renders). Freeze + 100%-refusal NOT live-exercised (STOP; red-proofs cover). NEXT: Peti onboarding
|
||||
runbook, NEW-box restore round-trip → old-box retirement.
|
||||
|
||||
- **2026-07-09 — SLICE 3 hub-verified escrow auto-confirm SHIPPED+LIVE (hub v0.40.0 + controller v0.108.0 +
|
||||
agent v0.79.0).** Replaces the operator-trust confirm with a verified fact: ceremony hashes the sealed
|
||||
repo password (agent, `restic_pw_sha256,omitempty` on the escrow upload — contract test moved in lockstep)
|
||||
|
||||
@@ -2,33 +2,36 @@
|
||||
|
||||
> **Overwrite** this file with a summary of the most recent task only (uniform with the other repos; not cumulative). The cumulative hub history lives in [hub/CHANGELOG.md](hub/CHANGELOG.md); the scripts history lives in [scripts/CHANGELOG.md](scripts/CHANGELOG.md).
|
||||
|
||||
## TASK — offsite SLICE 3: hub-verified escrow auto-confirm — 2026-07-09 — DEPLOYED
|
||||
## TASK — offsite SLICE 4: soft-quota + health monitoring + freeze lever — 2026-07-10 — DEPLOYED
|
||||
|
||||
**Shipped (felhom.eu):** hub **v0.40.0** `49d1233` + bump `c24d4af` — LIVE (Synced/Healthy). Chain pairs:
|
||||
agent **v0.79.0** (felhom-pve, 56/56) + controller **v0.108.0** (9201, healthy) — all live. Operator-free
|
||||
task (Viktor ran nothing).
|
||||
**Shipped (felhom.eu):** hub **v0.41.0** `fad5573` + bump `04826c9` — LIVE (Synced/Healthy, "Offsite
|
||||
checker initialized"). Pairs with controller **v0.109.1** (9201, healthy). The offsite arc's last build
|
||||
item; operator-free (Viktor ran nothing).
|
||||
|
||||
**The verification chain:** the ceremony hashes the staged repo password it seals (agent) → the hub stores
|
||||
`restic_pw_sha256` alongside the escrow blob (additive migration; NULL on legacy rows) and serves
|
||||
`escrow:{identity_blob_present, restic_pw_sha256, created_at}` in the **report ACK** → the controller flips
|
||||
offbox `EscrowState` pending→escrowed ONLY on `sha256(local repo_password)` match. **Blob-presence alone
|
||||
never confirms** — a stale blob (re-provision/inject/history) would be a false custody claim re-opening the
|
||||
fork-4 gap; this is the red-proofed core (blob-present-only check → stale-blob test flips → FAIL).
|
||||
Mismatch → pending + loud warn naming the ceremony (deduped per hash); never un-confirms. On flip the
|
||||
agent-staged secret is wiped (v0.107.0 path). The two hashers are pinned by the SAME cross-repo test vector
|
||||
(trimmed-string sha256). Manual confirm-escrow → documented deprecated fallback for legacy hash-less blobs.
|
||||
**Hub changes:**
|
||||
- **`monitor.OffsiteChecker`** — sibling of StorageFillChecker over the controller report's new `offsite`
|
||||
object: fill warn/crit at 90/95% of `quota_gb` (quota 0 = dedicated, silent) + `offsite_stale` (warning)
|
||||
for enabled+**escrowed** targets with no run >48h/never — the silently-stuck detector (recently-FAILING
|
||||
runs are `backup_failed`'s signal; pending/disabled never alert — **red-proofed**: dropping the
|
||||
escrowed-only filter made the pending customer alert → FAIL). Nil-safe on pre-v0.109 reports;
|
||||
same-second latest-report ties processed once per sweep.
|
||||
- **Freeze lever:** `Provisioner.SetOffsiteFrozen` — flips ONLY `readonly` on the exactly-1 labelled
|
||||
sub-account (SSH preserved — tested; ambiguity refuses); confirm-gated **Freeze/Unfreeze** buttons next
|
||||
to Re-issue (shared only). **MANUAL only, never automatic** — freezing also blocks prune, the customer's
|
||||
only way down from over-quota.
|
||||
|
||||
**Hub specifics:** `host_escrow.restic_pw_sha256` migration; `SaveHostEscrow`/`HostEscrow`/`GetHostEscrow`
|
||||
NULL-safe; `GetEscrowStatusForCustomer` (hosts⋈host_escrow, latest-updated wins); ACK object omitted when
|
||||
no escrow row (fresh customers stay silently pending); `TestEscrowUploadContract` mirrors agent v0.79.0.
|
||||
**Live verification (free checks, real pipeline):** the controller deploy's re-apply exercised
|
||||
**key-auth-first live** (re-pinned, zero password consumed) and surfaced a real bug — the re-apply demoted
|
||||
the escrowed demo to pending + wiped runtime status → **fixed as controller v0.109.1** (re-apply preserves
|
||||
custody + status; red-proofed), demo repaired (manual confirm — truthful, same escrowed password — + a
|
||||
run). The next report carried the full offsite object (`escrowed`, 45 MB / 50 GB, fresh run); the
|
||||
**checker held it healthy (0 offsite_* events)**; the `/backups` usage bar renders
|
||||
"Tárhelykeret: 43.2 MB / 50 GB (0%)".
|
||||
|
||||
**Live state check (no operator action):** migration applied on the production DB — all 3 legacy rows
|
||||
(demo included) read hash-NULL; the demo's `escrowed` target produced **zero** `escrow-confirm` controller
|
||||
log lines after reports through the new hub (Scenario E — never-revisit — proven live for free); no hub
|
||||
errors.
|
||||
**NOT live-exercised (per the STOP):** the freeze (would write-lock the demo's working offsite) and the
|
||||
100% refusal (demo is at 0.08%) — unit red-proofs cover both.
|
||||
|
||||
**NOT yet live-exercised:** the full happy chain (fresh enable → ceremony records the hash → ACK →
|
||||
auto-flip) — the first real offsite enable (Peti onboarding) proves it; unit red-proofs carry the logic.
|
||||
|
||||
**NEXT:** SLICE 4 (soft-quota on `quota_gb`), Peti onboarding runbook (live-exercises auto-confirm + F4
|
||||
re-issue), NEW-box restore round-trip → retire the old-box archive.
|
||||
**The offsite arc is now build-complete:** provisioning (S1/S2) → hardening (F1–F5, key-auth-first, wipe)
|
||||
→ auto-confirm (S3) → quota + health (S4). **NEXT:** the Peti onboarding runbook (live-exercises
|
||||
auto-confirm + F4 + fresh-provision quota end-to-end), NEW-box restore round-trip → retire the old-box
|
||||
archive.
|
||||
|
||||
+8
-1
@@ -94,7 +94,14 @@ the explicit recovery for a consumed-password dead-end — resets ONLY the resou
|
||||
stores it on `host_escrow` (NULL on legacy rows) and serves
|
||||
`escrow:{identity_blob_present, restic_pw_sha256, created_at}` in the **report ACK**; the controller
|
||||
auto-confirms its pending offbox escrow ONLY on a hash match with its current repo password (blob-presence
|
||||
alone never confirms). SLICE 4 (soft-quota) follows.
|
||||
alone never confirms).
|
||||
**SLICE 4 (v0.41.0) — offsite health monitoring + freeze lever:** `monitor.OffsiteChecker` reads the
|
||||
controller report's `offsite` object — fill alerts at 90/95% of `quota_gb` (quota 0 = dedicated, silent)
|
||||
and an `offsite_stale` warning when an enabled+escrowed target has no run in >48h (the silently-stuck
|
||||
detector; run FAILURES are `backup_failed`'s job; pending/disabled never alert). Nil-safe on pre-v0.109
|
||||
reports. The operator **Freeze/Unfreeze offsite** buttons (shared model only, next to Re-issue) flip ONLY
|
||||
`readonly` on the exactly-1 labelled sub-account — MANUAL only, never automatic (freezing also blocks
|
||||
prune, the customer's only way down from over-quota).
|
||||
|
||||
### Infrastructure Backup — RETIRED (Phase-1, 2026-06-16, hub v0.12.0)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user