- pbsdr: on a 403 pre-check (non-default storage id, no ACL yet) self-grant via the root wrapper then re-read, instead of aborting before the grant — closes F4/R-22. Red-proof TestSelfGrant_PreCheck403DoesNotAbortBeforeGrant.
- escrow preflight: late-bound CurrentPBSStorageID re-reads agent.json so a pbsdr-seeded pbs_storage_id flips the row green in-process (no restart). Red-proof TestEscrowPreflight_PBSStorageIDLiveReload.
- internal/poke: contentless UDP poke listener bound exclusively to the box WG /32 (port 51822), leading-edge debounced, fires the hub-loop out-of-band trigger for an immediate desired-state cycle. First slice of R-13. Red-proofs TestBindConfinement + TestDebounceCoalescesBurst.
Operator ruling 2026-07-13: walk the root's backing device through /sys/block/<dev>/slaves
recursively down to physical disks (dm AND md; topology, never VG names); those + any mounted-ESP
holder are system; the all-system fail-safe returns to being the WALK-FAILURE error case only.
SAFETY DIRECTION: a root-backing disk classified candidate is made impossible — per-branch
conservatism (any unresolvable slave fails the WHOLE walk -> ok=false -> the unchanged all-system
path).
- physicalDisksOf/walkSlaves in role.go (symlink canon -> wholeDiskOf fast path -> recursive
slaves walk; cycle/depth guard; non-/dev sources unwalkable)
- HostReader.BlockSlaves(name) — the ONE new seam method; ProcHostReader reads
/sys/block/<name>/slaves; all four test fakes mirror it
- role_walk_test.go: signature table (root-backing disk ALWAYS system across legacy-LVM /
md-raid / EFI+raw / EFI+LVM / nested dm-on-md — NEVER weaken) + dead-wizard-lives +
dangling-slave fail-safe (real sysKnown=false path) + cycle + empty-slaves; red-proofs A/B/D
run->fail->revert (recorded in REPORT)
- §3 spike transcripts (drill legacy: dm-1->sda3->sda; felhom-pve: ESP+walk agree on sda ->
byte-identical regression); caller audit: none relied on all-system as a feature
- format/mkfs paths, data-bearing guards, wizard UI untouched
Agent half of the verify-before-commit task (SPIKE-nas-verify-2026-07-11, b57f6c1):
retry=0 in the production NFS options (Q4-vi); ClassifyNetVerifyFailure on the live
Q4 strings (nfs_export merges not-found/not-permitted); add = sync fast-fail (2s TCP
pre-probe, nothing installed) + detached in-memory verify job judging /proc/mounts
only, auto-rollback on failure; GET /netstorage/verify-status (phase none = the
controller's Scenario-F rollback signal); unprivileged journalctl (systemd-journal
group, NO new sudoers grants).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
HashResticPassword = sha256 hex over the trimmed password (pinned
cross-repo vector). escrowUploadRequest gains restic_pw_sha256,omitempty
— set only when a staged password was sealed into the blob. Contract test
updated; hub mirrors next.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
The staged offsite repo password was wiped only by the escrow-create
ceremony; a confirm without a fresh ceremony (password already escrowed)
left the 0600 staged file behind. The controller calls this on every
EscrowState flip to escrowed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
The live scratch DR exposed the second half of PVE's all-or-nothing
explicit-params restore: mountpoints NOT named in the params are silently
DROPPED - the DR guest came up without its mp0/mp1 data volumes (boot
passed; the customer's world did not ride along). drRestoreOverrides now
derives the COMPLETE param set from the archive's extracted config:
explicit rootfs, every storage-backed mpN passed through (size + in-guest
path + backup flag preserved so vzrestore extracts its content), the two
structural binds replaced by 4d-swapped throwaways; unknown bind mpN or
unparseable size refuses loudly. Snapshot sections never shadow the
current config.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
Pool membership is what lets the pool-scoped token reach a guest; pct restore
--pool sets it only at CREATE, so a restore over an existing VMID drops the guest
from the felhom pool and 403s the next restore-test/DR on VM.Audit. This empty-pool
state is the true root cause of the campaign's "R1" (bind-mount restore failing was
a symptom — restore-test's existing bind neutralization never ran without config-read).
Add Client.PoolAddVMID (PUT /pools, additive+idempotent, Pool.Allocate) and call it
in bring-up after liveness when spec.Pool!="" — warn-not-fail on a hiccup (liveness
wins). B3 scratch-teardown 403 diagnosed as a cascade (restoretest already passes
Pool). Role/ACL untouched. Tests + red-proof.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
roleForMountPath resolved role only from the PVE storage view; a bind-mounted
RAW enrolled user-data drive is not a PVE storage, so it fail-safe'd to system
and the eject/decommission gates 403'd EVERY user-data drive in the standard
topology (campaign F2, where=/mnt/teszt_enroll role=system). Add a mount-table
fallback mirroring durableIDForMount Impl-2b: device-keyed classification with a
whole-disk containment pass (new storage.SameWholeDisk) and the Observe-error
early return kept BEFORE the fallback (else a blind view -> permissive). Only
roleForMountPath touched. Tests A1/B1/B2/C1-C3 + 3 red-proofs; existing RoleGated
tests green unmodified.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
Prerequisite for felhom-sshd (H1). Closes the SPIKE-felhom-sshd §8 lockout: a
second sshd's RuntimeDirectory=sshd removed the SHARED /run/sshd privsep dir and
took stock sshd on :22 down (sessions reset after KEXINIT).
Host artifacts (configs/, installed by felhom-host-install):
- felhom-privsep.tmpfiles: layer 1, boot-persistent /run/sshd owned by no unit
- felhom-mgmt-watchdog.sh/.service/.timer: layer 2, AGENT-INDEPENDENT ~60s heal
(stat-first recreate + reset-failed sshd only if failed + heal-marker); never
RuntimeDirectory=, never restarts stock sshd, never touches a healthy dir.
Go (internal/mgmtplane): read-only Reporter → additive omitempty mgmt_plane
heartbeat stanza (privsep_dir_ok/sshd_reachable/healed_recently/privsep_healed_at),
wired via Collector.SetMgmtPlaneReporter. Non-hollow tests + red-proofs.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
wgtunnel.InstallRecoveredKey: write an escrow-recovered WG private key (create-
only, refuse-overwrite) so the tunnel re-establishes with the same identity/pubkey
(same /32), no keygen. Wired into identity-consume -install-wg-key (opt-in;
pre-S3 blob → logged fresh-keygen fallback). Value never logged.
internal/dr (new): consume the host_loss restore_directive (was logged-ignored)
into an inspectable RestorePlan via the AddConsumer raw seam — per guest
{vmid,archive,target,sizing} + per drive {durable_id→mount} + offsite PBS coord.
DERIVE-AND-SURFACE only; the Consumer has no restore/destroy dependency (execute-
nothing is structural). guest_loss/absent → no plan.
Tests + red-proofs (WG create-only overwrite; plan mode-gate). No secrets on
argv/stdout/logs. The destructive in-place restore is a separate operator-present
STOP-gated drill.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
A WAN (pbs-tier) restore of a large guest exceeds the restore-task wait's 10m
default → the wait expired mid-restore, teardown fired against a still-restoring
(not-yet-pool-associated) scratch guest → leak + a phantom VM.Allocate 403.
- RestoreTestSpec.RestoreTaskTimeout (0→10m default); the restore WaitTask passes
it. Local tier unchanged (10m).
- config RestoreTestPBSRestoreTimeoutSeconds + accessor (default 120m).
- main restoreTaskTimeout(cfg,tier): configured PBS timeout only when tier==pbs,
else 0. Both scheduler + selftest spec builds.
- Tests + WaitOptions red-proof + accessor contract.
The "grant scratch-band VM.Allocate" follow-up is diagnosed not blind-applied:
the scratch is restored INTO /pool/felhom (ACL already grants VM.Allocate), so
the earlier 403 was a consequence of the timeout. No ACL/host-install change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
Phase-1 live probe (felhom-hetzner) proved backup/restore/list/isolation over
the tunnel with a per-customer DatastoreBackup token, but the agent's PBS client
was namespace-unaware: Snapshots hit the datastore root (403 for a scoped token)
and Verify was whole-datastore (needs Datastore.Verify ~ admin). Operator-
approved fix.
- pbs.Config.Namespace + Client.namespace; Snapshots appends ?ns=; Verify sends
ns= (ns-scoped verify works with DatastoreBackup on the own ns — no admin
widening, Phase-1 confirmed). Root-ns clients unchanged (whole-datastore).
- proxmox.Storage.Namespace (parsed from /storage `namespace`).
- pbsTargetsFromPVE threads s.Namespace into the client.
Confirmed tenant ACL: DatastoreBackup on /datastore/felhom-offsite/<ns> (NOT
/ns/<ns>) to BOTH felhom@pbs (user) AND felhom@pbs!<ns> (token) — PBS privsep =
intersection; isolation holds (cross-ns 403 proven). TestClient_NamespaceScoping
red-proofed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
v4-pin (doc 06 §4.2): renderConf takes a pre-resolved IPv4 literal and writes
Endpoint=<ip>:<port> — never the DNS name, never AAAA. Resolver seam (A records
only, LookupNetIP "ip4"); multiple A → lowest (deterministic fleet-wide);
renderConf stays pure. Resolved IP cached: steady-state Apply = zero DNS + zero
execs. DNS failure keeps the last conf (never a teardown).
Watchdog (loop-only, so Apply's zero-exec steady state is untouched): handshake
age > wg_tunnel.stale_after_seconds (default 180) → re-resolve; IP changed →
re-render + restart (endpoint re-IP recovery); IP same → no churn (throttled
warn). Staleness read reuses wg show latest-handshakes (never dump).
Capability: wg-conf-install/enable/restart/handshake-read flipped Critical=true
(backups ride the tunnel from S4); apt-install + disable stay non-critical.
TestWGCapabilityCriticality pins the set.
Tests + red-proofs a/b/d all fire. No new sudoers grant; no wire/JSON change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
The 2026-07-04 CGNAT smoke test found MTU 1420 silently black-holes bulk TCP
on sub-~1480 paths (mobile ~1400, DS-Lite ~1452): handshake+ping stay healthy,
PBS TLS page (and at S4 the backup itself) drops. Set a fleet-wide, permanent,
family-agnostic client MTU of 1280 (RFC 8200 IPv6-minimum floor; outer 1340 v4
/ 1360 v6 fits every realistic path).
Client-only by construction — interface MTU caps box→PBS, advertised MSS caps
PBS→box; the endpoint's wg0 is untouched (zero live-endpoint risk). New const
clientMTU=1280 as the single home; golden pins exact "MTU = 1280" (red-proofed
against a 1420 flip). Stale report.go comment updated. No wire/JSON change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
B3: Lookup re-reads the append-only store once on a miss (cross-process
coherence with the one-shot provisioner; size short-circuit bounds the cost;
behind the TokenAuthority seam). B2: fenced mkdir -p /var/lib/vz/snippets
before the snippet install + the one narrow sudoers grant. Both red-proofed;
drill findings DRILL-day0-cleanroom-2026-07-03 B3/B2.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
staleLockController.Guests() = ListLXC ∩ GET /pools/felhom members (ownership
PROVEN via the pool registry, never assumed from enumeration scope); pool-read
failure fail-safes the whole recovery through the existing guest-list guard.
New Client.Pool read (needs Pool.Audit — host-install v1.9.0; Pool.Allocate
does NOT satisfy it, spike T2). Composed pve:pool-read capability (non-critical)
+ --selftest pool-read line. Red-proofed negative tests drive the REAL
controller over a broad-token-shaped fake.
Per SPIKE-a1-pool-membership-read-2026-07-03.md; audit A1
(AUDIT-blast-radius-hostroot-localapi-2026-07-02).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
From AUDIT-blast-radius-hostroot-localapi-2026-07-02.md. Each fix ships with a
non-hollow test + a companion red-proof (shown failing on the pre-fix impl).
Sudoers install-source grants became globs — deploy the sudoers drop-in with
the binary. A1 (stale-lock pool-membership) deliberately excluded (spike).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PSK5g6qYLknKj8u3QAFEr6
Campaign pool-effects F1 (HIGH): the bring-up compensating rollback and the
restore-test teardown destroyed the target vmid even when RestoreLXC failed
synchronously without creating anything — destroying a guest the transaction
never made (only the pool ACL 403 contained it). A RestoreLXC UPID is now the
sole destroy authorization in all three destroy paths (in-process bring-up
defer, in-process restore-test teardown, Recover). F2: the restore-test
advances past an 'already exists' band vmid (invisible squatter) instead of
failing + false-alerting; a fully-occupied band Skips.
Red-proof verified: with the gates reverted, the four new tests fail with the
innocent-guest destroy. go build/vet/test clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Agent-view showed "—" device + no size for a raw (no-PVE-storage) drive because
the registry union row never set backing_device/total_bytes/used_bytes (Observe
drives get those from pvesm status). Resolve BackingDevice via ByUUIDDevicePath +
read capacity via statfsCapacity (build-tagged syscall.Statfs; no-op off-Linux).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The Impl-2a registry union row omitted GuestPath + BoundUnderParent, so the
controller read a registry-only (raw) drive as "Leválasztva" despite being mounted
+ bound live. Populate them like the Observe path (StablePathForRaw +
boundUnderParent). Last piece of first-class raw-drive support.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
buildLocalAPIServer never passed Options.HostReader, so s.host was nil — the
v0.56.0 durableIDForMount raw-mount fallback + the role gate's host classification
silently no-op'd. Wire storage.NewProcHostReader(). This makes the raw-drive
intent/guest-bind resolution actually fire live.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ReassertGuestBinds mapped durable-id->mount from Observe() only, so a raw enrolled
drive was never found and its in-guest bind wasn't re-asserted after reboot/re-mount
(drive showed "Leválasztva"). Augment the map from the mount table (raw /mnt/<name>
-> device fs-UUID via HostReader), skipping the /mnt/felhom-drives bind; Observe
failure no longer aborts. Completes v0.56.0's raw-drive durability.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
durableIDForMount was Observe-only, so a raw (non-PVE-storage) drive's enroll
intent + guest-bind went unrecorded ("durable-id unresolved") — it mounted+bound
but wasn't intent-tracked (RegistryKnownTargets skips intent==new). Fall back to
resolving the mount's device fs-UUID (HostReader.Mounts + ResolveUUID) →
uuid:<fs-uuid>, same scheme as Observe. Test + red-proof. Residual:
ReassertGuestBinds still Observe-based (raw guest-bind not re-asserted post-reboot).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
GET /disks/candidates enumerates host disks the Impl-1 unclaimed filter proves
free (init/attach split). RegistryKnownTargets sources the watchdog's known-drive
set from the intent registry + Felhom .mount units (not Observe/PVE storages) —
decouples drive health from PVE storage (closes the registry-only false-detach
class); Observe kept for real PVE storages + a deduped /disks union. Idempotent
existing-drive migration at start. Tests + red-proof (Observe misses a
registry-only drive; registry provider tracks it). go build/vet/test clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Impl-1. Format now runs a mandatory unclaimed-disk guard (internal/storage/claim.go:
SystemDisks + lsblk member-FSTYPE + foreign-mount + RO + pvs/zpool; fail-safe →
CLAIMED) before any mkfs — refuses the OS disk / LVM PV / ZFS-mdraid member /
foreign-mounted device even when non-data-bearing (guard sits in Format, not the
handler). Below the agent, mkfs goes ONLY through configs/felhom-mkfs-guarded.sh
(sudoers no longer allowlists raw mkfs.*), which re-checks the catastrophic cases
as root. Read-only pvs/zpool added to FELHOM_DISK. Tests + red-proof; capability
manifest updated. go build/vet/test clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
RestoreLXCOptions.Pool → pct restore --pool (omit-when-empty). New
reconcile.DefaultPool="felhom"; BringUpSpec.Pool threaded to the bring-up
restore; BOTH restore sites pool the guest (provision/DR via spec.Pool set to
DefaultPool by the CLI; restore-test scratch via DefaultPool = SPIKE residual
#2). No agent ACL change (ships in host-install v1.6.0); the pool param is inert
until the token has Pool.Allocate + the pool exists, so publishing is safe ahead
of the coordinated swap. Tests + red-proofs; go build/vet/test clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
New -cores N / -memory M (MiB) flags for --selftest=bring-up|provision (0 = keep
golden default), flowed through bringUpSizing into reconcile.BringUpSpec so the
existing buildBringUpConfig emits cores/memory into the pre-start config PUT (the
cap lands before first boot). No engine change. New pure-function test
TestBuildBringUpConfig_ResourceCaps asserts the set + omit-when-zero cases
(red-proof run + reverted). go build/vet/test clean.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The preventive counterpart to host_disk + storage_fill detectors: the periodic local
whole-guest vzdump now prunes its own old archives (keep-last=3, clamped >=1) so a box
can't refill its own root via its own backups. Local target only — PBS never pruned
(resolved via ListStorage; fail-safe skip on unknown). Seeded in host-install.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HxLA1mZurFq9kt8hneFeCs
Host-side NFS/SMB automount of a bulk-media NAS share under /mnt/felhom-drives/<name>
(propagates into the guest via the existing shared bind), the +100000 uid recipe,
per-share liveness, and add/list/remove local-API endpoints. A NAS is a distinct class
that bypasses the drive enroll/eject/decommission/SMART/watchdog machinery.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HxLA1mZurFq9kt8hneFeCs
F2-a red-proof PASS: parent came up own peer group (shared:94, not
shared:1), no doubling, guest sees both drives. F2-b PASS: reboot with the
exact snapshot-delete stuck state → agent auto-recovered (unlock → removed
real dangling vzdump snapshot → started CT). Zero spurious pages. Two
live-caught fixes noted (?source=active, AllowWarnings).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0162BnMpUXscPsUB1cU8Tr6K
F2-b: at startup, recover a guest left with a stale vzdump lock by a
reboot-during-backup — pct unlock -> delete dangling vzdump snapshot ->
start iff onboot, guarded by a no-vzdump-running invariant (fail-safe).
New internal/localapi/stalelock.go; proxmox GuestConfig.Lock()/OnBoot(),
ListSnapshots, ListRunningTasks, Snapshot type. New narrow sudoers grant
FELHOM_STALELOCK (pct unlock) + Critical capability stalelock-unlock.
F2-a: EnsureSharedParent only redeployed the boot script when the UNIT
differed, so the v0.36.6 make-private fix never reached hosts whose unit
was current -> /mnt/felhom-drives stayed in root's shared:1 and doubled
every drive bind. New sharedParentInstallStale compares BOTH script and
unit. Boot-time-only; never churns the live mount.
Both root causes confirmed live on felhom-pve before fixing. Green gate
(build/vet/test) all pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0162BnMpUXscPsUB1cU8Tr6K
HostReport.LeafFingerprint rides the served fp (from EnsureLeaf) on every report; empty when local
API disabled. Collector.SetLeafFingerprint threads it like Capabilities. Hub watches it for a re-key.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pg8ANF97SEeKYSN5Jxw3qJ
controllerHealthy reads RestartCount (running&&rc>0 -> not ok) + signals needsDwell for no-healthcheck;
verify requires verifyDwell(=3) consecutive ok polls for a no-healthcheck image (real healthcheck
trusted immediately). Closes the F1 hole (alpine crash-loop passed the point-in-time check). Red-proof
+ dwell + real-image tests. No sudoers/orchestration change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Pg8ANF97SEeKYSN5Jxw3qJ
admin