R-897 filed: after a host restart the guest's first backup run can race the agent's drive re-bind (measured on demo-hp, kernel-lane arc)

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0159rPz1ZhFKsS53msqPYxtS

documentation/backlog/OPEN-ITEMS.md

@@ -164,6 +164,7 @@ stopping line that lies.
 | **R-540** | Backup & restore | P3 | **[P3-LOW] The hub knows exactly ONE off-site pool box, so there is no rule for what happens when it fills.** Read from source 2026-09-16 while making off-site the default: `HETZNER_POOL_BOX_ID` is a single value, and every shared customer becomes a sub-account on that box. With off-site now ON for every new customer (hub v0.116.0) the box fills faster, and the fill warning (80%/90% of the box, `monitor/offsite.go`) tells the operator it is filling but nothing says which box a new customer should land on. **Needs a selection rule** (least-full, or explicit per-customer), not a bigger box. No customer is at risk today: the pool box read 0.3% full (2.7 GB of 1 TB), Σ shared quota 150 GB, oversub 0.15x. | **READY — rank P3-LOW; owner: CC (hub)** **2026-10-05 (burn-down night): NEEDS A DESIGN (and money).** A selection rule needs a multi-box configuration and a second box. Next: an operator pick of the rule and its trigger (e.g. add box 2 at 70 %). | — | — | CC |
 | **R-548** | Backup & restore | P3 | **[P3-LOW] The whole-guest backup’s LOCAL tier cannot fit on a small-system-disk box, and will retry on that tier for ever.** MEASURED 2026-09-17 (chaos night) on `tester-1-022354`: a whole-guest backup wrote a **~29 GB source** (`mp0` = `local-lvm:vm-9201-disk-1`, 70 G provisioned, 40.58 % used, `backup=1`) into `pve-root`, which on a 32 GB system disk is **14 GB total with ~4.9 GB free**. Two samples thirty seconds apart showed the archive growing ~497 MB while free space fell ~475 MB — **~16 MB/s, i.e. under four minutes to a full `/`** on the nested PVE. **The product’s behaviour is correct and legible throughout:** it failed the tier and said which one — `whole_guest_backup_failed` (error, operator-only): „Whole-guest backup FAILED on the local tier — retrying with backoff (next attempt in 15m0s)” — its status surface agreed (`target_id:"local"`, `success:false`, `size_bytes:0`), and the **off-site tier then ran from the same snapshot and succeeded in ~8½ minutes, encrypted to ep0, consuming no local disk at all**. So the data still left the house. What is filed is the loop: on a box shaped like this the local tier can **never** succeed, and it keeps retrying on a backoff for ever, burning I/O and risking `/` each time. **Honest caveat:** the 32 GB system disk is this drill’s fixture choice, so the row is conditional on disk size — but nothing in the product checks whether the local target could ever hold the source before trying. **Fix shape:** compare the source size against the target’s free space before starting the local tier and skip it with a clear reason, rather than discovering it at ~16 MB/s. Evidence: `audits/evidence-chaos-night-2026-09-17/round-6.txt`. **-- 2026-09-30, the class on a real box: demo-hp's whole-guest LOCAL tier was refused by the space preflight (R-685) 10 times since 2026-09-27** („local has 12.1 GiB free; the last archive of guest 9201 was 8.9 GiB, so a new one needs about 12.1 GiB") — its newest local archive was 2026-09-29 04:42, 30 h old at the day's read (the weekly PBS tier, last 2026-09-24, is its whole copy meanwhile). The refusal is correct and named; what is missing is that nothing gives the box the room back (the host's `local` holds the 8.9 GiB archive of the only guest plus templates on a 39 GB root). `audits/pg-last-six-2026-09-30/C/`. | **READY — rank P3-LOW; owner: CC** | — | — | CC |
 | **R-698** | Backup & restore | P3 | **[P3-LOW] A backup stores the image's NAME, not the image — a restore of a version its maker has deleted cannot start.** `RecoveryManifest.image_pins` ("image NOT stored — re-pulled on restore"); since controller v0.275.0 each data file also records its running `ref@digest`, and a restore brings the data back AT ITS OWN VERSION (`07` §6.6) — so a restore asks for exactly the old image. **Measured 2026-09-26** (`audits/version-travel-2026-09-26/A7/`, registry HEADs, no pulls): the catalog's 42 ladder `ref@digest` pairs all resolve (200); an invented digest answers 404 on Docker Hub and ghcr.io (negative control). Not measured: the digests recorded on boxes (older than any ladder entry), how often makers delete versions, the catalog's 66 digest-less compose lines. **Options (decide nothing yet):** (a) keep — a restore of a deleted version fails at the pull and the household uses the next copy or a newer version; (b) mirror every INSTALLED image into the DooPlex registry, restore falls back to it — storage + bandwidth on DooPlex, a new part on the recovery path; (c) mirror only ladder-named versions — bounded, misses pre-ladder boxes; (d) `docker save` into the unit — hundreds of MB per app per copy on every tier. **-- 2026-09-30 late (decision 53):** a box now keeps only an app's running and previous image; a restore to an older version re-pulls it — as every restore already did. The limit above is unchanged. | **OPEN — P3; owner: operator (a decision), CC measures** **Operator ruling 2026-10-05 18:23: kept OPEN as a known risk to a household's restore; owner the operator; not worked on in the burn-down.** | — | — | operator |
+| **R-897** | Backup & restore | P3 | **After a host restart, the guest's first backup run can start while the agent re-binds the drive, and that app's backup fails.** MEASURED 2026-10-07 on demo-hp (a plain reboot for the kernel lane's wait, `audits/kernel-lane-2026-10-07/B/demo-hp/`): the agent bound `/mnt/felhom-drives/hdd_1` at 15:16:07 (prior_binds=0), the guest started after it and could not see that bind, so at 15:16:28 the reconcile NORMALIZED it (umount + mount, prior_binds=1, `internal/localapi/intermediary.go`); in that same second the controller's backup refresh failed for calibre-web: `recovery_unit_capture_failed` — `mkdir /mnt/felhom-drives/hdd_1/backups: permission denied` — and `backup_run_failures` (1 of 9). Every later 20 s tick was quiet; the app stayed healthy. **Why it matters now:** the kernel lane restarts a box at night (`09` §3 decision 172), so each kernel step can raise a false backup failure the morning after. Fix direction (a design, not taken here): the controller's post-boot backup run waits until the drive reads bound-and-live (`BoundUnderParent`), or the agent's first bind waits for the guest to run so no normalize is needed. | **OPEN — filed 2026-10-07 (kernel-lane arc).** | — | — | CC |
 | **R-91** | Backup & restore | P4 | Old 13 GB datastore copy at `/srv/pbs-felhom` on ep0's root disk **Checked from source 2026-10-05 (burn-down round 2):** Gate is long past (row waits on demo-felhom's first post-migration PBS backup, migration 2026-07-27). Last positive record of the copy: audits/CAMPAIGN-9-restore-proof-2026-07-28.md:759 'ep0 : /srv/pbs-felhom rollback copy intact (13G)'; CONTEXT.md:3666 still says it is 13 G of dead weight awaiting R-91. No later record of deletion found (grep srv/pbs-felhom across felhom.eu). Deleting is on ep0 (protected) and needs an operator word. | WATCHING | demo-felhom's first **post-migration** PBS backup | Delete once it lands; fix `CONTEXT.md:1018` same commit | CC |
 | **R-164** | Backup & restore | P4 | **C2's chain: the DB volume tar cannot be dropped until a SOUND dump predicate exists.** The unit carries both a volume tar and a SQL dump; the restore uses **both** — the dump is authoritative and replayed *after* the tar so it WINS (F17), with only the DB service up (R-47) — `internal/backup/restore_unit.go:262-266`. Dropping the DB container's tar would halve DB-app units **and** close the R-127(b) initdb-skip password trap (restored PGDATA ⇒ `POSTGRES_PASSWORD` ignored). | **BLOCKED** — on the predicate | a dump-validity predicate that is not `accounts has rows` | **The obvious gate is DEAD, measured:** `ValidateDump` warns when the `accounts` table is empty, and that warning was **correct** — the live DB genuinely had 0 accounts, and seeding one stopped the warning and put the row in the dump. But **a fresh appliance legitimately has zero accounts**, so promoting that predicate to a gate would **block every new customer's first backup**. Order: (1) a sound predicate — dump vs **live** per-table counts, not an absolute expectation; (2) warn→gate; (3) tar-drop. **Until (1), the tar is load-bearing** — not because dumps are bad, but because nothing can yet prove one is good. Pairs with **R-127** | CC |
 | **R-213** | Backup & restore | P4 | **Putting files back in place — the half the recovery screen deliberately does not do.** The screen (R-193, controller v0.200.0) unlocks the repository and LISTS what is in it; restoring is per-app and lives in the backups area, and the operator ruled the two separate on 2026-08-05: *a screen that unlocks and then offers to overwrite is two decisions wearing one button*. What is missing is the step after the listing — a customer who can now SEE their files still has to work out, per app, which restore to choose. **The operator named its requirement: a live-versus-backup comparison** — the customer must be able to see what would change before anything is overwritten | **OPEN — not started, deliberately** | the comparison design (nothing exists for it yet) | Design the live-vs-backup comparison, then the put-back flow on top of it. Do NOT fold it into the recovery screen | Operator + CC |