diff --git a/REPORT.md b/REPORT.md index 49e10e7..9fa07b9 100644 --- a/REPORT.md +++ b/REPORT.md @@ -51,7 +51,7 @@ placement stays non-auto-deploying; no engine changes beyond `offbox_restore.go` ## Commits - felhom.eu hub: `08fef48` -- felhom-controller: `` +- felhom-controller: `0cfcc42` ## Tests — results @@ -77,15 +77,47 @@ All reverted; post-revert both suites green; no mutation residue. - **Hub:** built + pushed `felhom-hub:0.55.0` on 180; `manifests/hub.yaml` → :0.55.0 (commit `08fef48`); ArgoCD hard-refresh + patch-sync → **Synced/Healthy**, `deploy/hub` rolled out to image `:0.55.0`, startup `Listening on :8080`. -- **Controller:** `` +- **Controller:** built + pushed `felhom-controller:0.134.1` on 180; deployed to guest 9201 → + `Up (healthy)`. Commit `0cfcc42`. ## §13 live validation -`` +**Credential blocker (honest):** the sanctioned credential `180:~/.config/credentials` `C4_PASSWORD` +(len 14) is **stale** — the demo customer changed their dashboard password after the one-time claim +(login returns `Hibás jelszó`, no session). So the dashboard-session legs (manual run trigger, the two +scratch-restore legs, the UI prefs round-trip, the synthetic delivery event) could NOT be driven. I +did **not** reset the customer's password or extract the controller's hub API key to force a session +(both would alter demo state / overreach). What was validated read-only / organically instead: -## NOT yet live-validated — awaiting CAMPAIGN-6D (supervised) +- **Leg 2a — enlarged snapshot shape (LIVE, organic):** the daily 04:15 CEST scheduled run (2026-07-15 + 02:15 UTC, v0.134.0) produced the multi-path shape. `restic snapshots --no-lock --json` newest per + tag: **calibre-web** = `["…/nas-media/backups/primary/calibre-web", "…/nas-media/userdata/media/books"]` + (unit + MANDATORY userdata); the 2026-07-14 calibre-web snapshot was unit-only — the shape change is + visible across runs. immich / audiobookshelf newest = unit-only (no resolvable mandatory HDD bind). +- **Leg 2b — raw-data quota (LIVE):** stored `repo_size_bytes = 280,932,901` (267.9 MB) == live + `stats --mode raw-data` (280,932,901) exactly — the raw-data switch refreshed the stored value (was + 744,763,144 / 710 MB modeless before 3a). 15 snapshots retained. +- **Leg 2c — forget grouping:** in the deployed code (both call sites) + unit-tested; the live log line + was not capturable (the container restarted at the v0.134.1 deploy, rotating the 04:15 run's logs). +- **Delivery chain (partial, read-only):** hub v0.55.0 allowlist entry is LIVE (deployed Synced/Healthy + + the hub acceptance test). Controller notifier enabled (`Notifier enabled (hub: https://hub.felhom.eu)`). + The prefs migration is confirmed getter-only (stored `enabled_events` does NOT persist + `offbox_enlarge_blocked` — matches the design; surfaced at read). +- **Hygiene:** no scratch dirs were created (legs 3–4 did not run); the login-attempt temp files on 180 + were removed; read-only inspection scripts removed from the container. -`` +## NOT yet live-validated — awaiting a supervised session with the CURRENT customer password / 6D + +- Manual run trigger via the UI endpoint (organically covered by the scheduled run for shape + size). +- Unit-only + full scratch restore via the real endpoints (legs 3–4). +- The prefs checkbox round-trip in the live UI (leg 5) — migration + sync are unit-tested + getter + confirmed live; the checked-render + save-survives round-trip needs the session. +- The synthetic `offbox_enlarge_blocked` delivery event + customer email (leg 6) — hub ingestion is + live; end-to-end delivery needs a session or the demo API key. +- `PlaceOffsiteRestore` against live data; organic enlarge-block firing (demo repo 268 MB / quota 50 GB); + the SQ3 immich offsite-only full circle. +- **Follow-up flagged:** refresh `C4_PASSWORD` in `180:~/.config/credentials` (the demo customer's + current dashboard password) so future in-session live legs aren't blocked. ## Observations (documented, not acted on)