CLAUDE.md — a new "How a deploy reaches that address space" block under the v3
Batch 4 address-space section. The blanket ChangedRawTags/ChangedUnsReferenceVariables
=> Rebuild rule is gone, so the doc now names both halves explicitly: what still
rebuilds (alarm-bearing raw-tag deltas, raw-tag identity changes, raw container
renames/removals, changed equipment/alarms) and what is now surgical (non-alarm
attribute deltas, every UnsReferenceDelta). Also records the three semantics that
are easy to break by accident: mirror repair rides UnsReferenceVariable's inherited
backing fields (omitting a newly-inherited field silently reopens the staleness),
a degraded non-rebuild apply deliberately does not advance _lastApplied, and the
raw-realm announce gap is no longer neutral now that surgical adds keep
subscriptions alive.
CLAUDE.md ConfigSource — the RevisionHash == SHA-256(artifact bytes) paragraph is
still true and stays; it gains the clause that the blob is gzip-compressed since
2026-08-15 and the hash covers the COMPRESSED bytes, which is precisely why the
fetcher, cache and serve path were untouched.
docs/Uns.md + docs/Raw.md — one section each on what a deploy now tears down, from
the authoring surface's point of view, including the mirror repair semantics
(identity-preserving backing edits update in place and keep their MonitoredItems;
re-points and effective-name edits remove+recreate that one node).
docs/Configuration.md — "Artifact compression (2026-08-15)" under ConfigSource/
ConfigServe: gzip inside the hashed bytes, magic-sniffed on read, pre-upgrade
uncompressed rows still parse, plus the two operational facts an operator needs.
The first deploy after upgrade re-hashes every revision (same config, new bytes,
new RevisionHash) so it is a one-time full re-dispatch rather than a no-op; and
rollout is ordered — nodes must run artifact-sniffing code BEFORE central starts
emitting compressed blobs, since a pre-upgrade node handed a gzip artifact fails
the apply and keeps last-known-good until it is upgraded.
Plan doc — a "Follow-ups (recorded, not implemented)" section covering the seven
items surfaced during implementation and review: the raw-realm announce gap, the
RemoveEquipmentSubtree dangling-inverse-edge exposure, the missing end-to-end test
for a degraded remove-bearing apply's ratcheted rebuild, the scripted-alarm
thin-plan-ids-against-rich-rows blindness, the absent monotonic dispatch sequence,
LocalDb store ordering under concurrent StoreAsync, and threading the in-hand blob
through RestoreApplied.
tasks.json — all nine tasks completed.
Docs only; no source changes.
- docs/v2/Galaxy.Performance.md gains a "Value-path batching (2026-08-15)"
section: the batch seams (IBatchSubscribable / PollGroupEngine onBatch ->
AttributeValueBatchPublished -> AttributeValueBatchUpdate -> sink WriteValues
with 512-chunked single-lock batches in the node manager), the invariants (#519
status codes verbatim, per-ITEM metering, Modbus deadband honored via the
driver-filtered buffer, the ValueWriteBatchException partial-failure contract),
the RealmKey re-key, and the pinned dispatcher's throughput 1 -> 100.
- The plan doc gains a "Follow-ups (recorded, not implemented)" section: the
PollGroupEngine unconditional-collection trap (a future per-tag filter in
S7/AbCip/FOCAS would silently reintroduce the deadband-bypass bug), a two-level
driverId -> (rawPath -> refs) routing map, null/empty guard consistency on the
internal batch messages, a possible DriverHostActorLiveValueTests merge, and the
pre-existing aborted-cycle tail-loss on the three unbuffered drivers.
- tasks.json: all 13 tasks completed.
CLAUDE.md needed no edit: its only statement in this area ("One driver publish for
a RawPath fans (in DriverHostActor) to the raw NodeId AND every referencing UNS
NodeId") is still true — batching changed the message granularity, not the
fan-out semantics.
Verification: dotnet build ZB.MOM.WW.OtOpcUa.slnx clean (0 errors, no new warning
classes, zero warnings in any production file). 9 unit projects green, 3470
passed / 5 skipped / 0 failed: Core.Abstractions 305, Core 351+1s, Commons 348,
Runtime 575+3s, OpcUaServer 424+1s, Modbus 399, S7 302, AbCip 410, FOCAS 356.
No DB-backed suite was run — tests/Server/ZB.MOM.WW.OtOpcUa.Server.Tests does not
exist, and nothing this branch touches is covered by the SQL-backed projects.
Blazor does not serialize OnParametersSetAsync against its own outstanding
task, and the modal's `_open` bool cannot tell two opens apart. ResetAndResolveAsync
awaits LoadMergedProbeConfigAsync and ResolveGroupContextAsync and then wrote
_driverType/_mergedConfig/_effectiveDeviceId/_groupPrefix/_canBrowse/_disabledReason
with no staleness re-check, so a close-reopen onto another device let the first
open's late continuation install its device's browse context under the displayed
modal — "Connect & browse" would silently dial the wrong controller and the commit
would land under the wrong tag-group.
Ports the per-open `_openGeneration` counter proven in the two sibling modals
(RawManualTagEntryModal, RawTagModal, 2026-08-05) to every post-await write here:
after both awaits, in the catch's error banner, and in the finally. The finally is
the one departure from the sibling shape and it is load-bearing — this modal resolves
inside a try/finally, so a stale `return` still runs the tail and would drop the LIVE
open's spinner mid-resolve.
Pinned by four RawBrowseModalStaleOpenTests facts, each observed red before the guard
and each re-falsified afterwards by reverting only its own guard. The modal calls
StateHasChanged on the open path, which throws without a render handle, so the test
attaches an inert Renderer subclass (file-local BL0006 pragma); [Parameter]s are set
by reflection per the BL0005 convention.
AdminUI suite 1117/1117 green.
The connect path was the one capability call the resilience pipeline never
saw, so a driver that could not connect contributed zero breaker input and
the /hosts resilience panel read 'All targets healthy' against a dead
endpoint. InitializeAsync now routes through the invoker as
DriverCapability.Initialize, keyed on the driver-instance id.
- No pipeline retry: the actor's RetryConnect timer is the retry loop
(parser forces retryCount overrides to 0, same funnel as write-shaped).
- Breaker threshold 2, not the Read-sized 5: the breaker samples 30 s and
the reconnect timer fires every 10 s, so at most three attempts fit one
window — threshold 5 could never open. Pinned by its own test.
- An open breaker fast-fails into the same InitializeFailed path as a real
connect failure; the 15 s break window costs at most one extra tick.
The best-effort teardown ShutdownAsync stays unwrapped.
- Companion fix: DriverCapabilityInvokerFactory.Create now invalidates the
instance's tracker snapshots with its pipelines — a respawn during an
outage discarded the only breaker whose OnClosed could clear the OPEN
flag, stranding a false 'Breaker OPEN' chip forever.
Live gate PASSED on docker-dev (image symbol-verified per container):
rtu-gate against dead 10.100.0.35:5021 opened its Initialize breaker on
both centrals with the red 'BREAKER OPEN' chip on /hosts; a socat
5021→5020 forwarder recovered it (both breakers closed within 1 s, panel
back to 'All targets healthy'); healthy MAIN-modbus recorded zero breaker
events. Full suite 50/50 green. Evidence:
docs/plans/2026-08-07-adminui-followups-and-live-gates.md §7.
The two items deferment.md §0.1 named as the remaining untracked engineering.
adminui-followups (19 pending / 0 completed) — CLOSED by the overlap audit the plan
itself asks for first, with no code written. All 19 tasks are shipped, superseded by
v3, or moot. WS1's premise had also expired: it assumed driver-embedded tag lists in
DriverConfig are the runtime source of truth, which v3 ended (the artifact populates
RawTags). Its capability shipped in the v3 shape — TagConfigEditorMap + DeviceFormMap,
both verified reachable from RawTagModal/DeviceModal rather than merely present.
Live gates on docker-dev (image rebuilt: the rig was running a build 3 h older than the
last code commit; parity asserted per container plus a positive check that the symbols
under test are in the shipped DLLs):
- R2-03 T11 PASSED — build 0 errors, Runtime.Tests 555/555, RoslynVirtualTagEvaluator 20/20.
- R2-03 T12 PASSED — VT 42 -> 0x80020000 -> 42, staged WITHOUT a deployment (a Modbus
register flipped on the fixture) so the VT child survived; a deploy-driven trip creates
a fresh child with no last value, which is correct-by-design and masked the result twice.
P7 confirmed: two byte-identical redeploys, zero disturbance.
- driver-reconfigure Task 3 PASSED — a faulted driver adopted a corrected config and
connected with the node's startedAt unchanged and RestartCount=0.
- R2-10 T10 still OPEN, but the recorded cause was wrong. Not ConnectionBackoff
throttling (the breaker opens on 2 failures / 30 s): the dead-endpoint failure mode
never enters the resilience pipeline. IDriver.InitializeAsync is not invoker-wrapped
(DriverInstanceActor.cs:707) and ModbusDriver swallows per-tag failures in BOTH
directions, so even the plan's fallback trigger cannot work. Live-proven: a driver
timing out every read against a blackhole reported ConsecutiveFailures=0 while the
panel read "All targets healthy."
Finding from T12: the Bad publish's documented "carrying the last-known value" is
unobservable by any client — subscription and a raw Read service call both return NULL.
The actor is correct; every existing failure assertion checked StatusCode alone, so the
value half was asserted nowhere. Adds
VirtualTagActorTests.The_Bad_result_carries_the_last_known_value, which proves the actor
emits it. The documentation is what overstates.
Re-gates Akka peer-to-peer secret replication on the current three-mesh docker-dev
rig and supersedes the 2026-07-18 six-node gate (lmxopcua#482, c878fbbd) as evidence.
That gate proved convergence to all five peers on the old single fleet-wide mesh;
after mesh Phase 6 (2839deb5) partitioned the rig into three independent 2-node
meshes, six-node convergence would be a FAILURE, not a pass. Phase 6's own leg 6
recorded "PASS (by fallback criterion)" — it inferred pair-locality structurally and
never seeded a secret. This replaces that inference with direct evidence.
All four checks PASS, on master @ 21009874 with one image (OtOpcUa.Host.dll
a434b04f byte-identical on all six nodes), Secrets libs 0.2.3:
1. Pair convergence — seeded on site-a-1 alone, on site-a-2 in <=7s (budget 60s),
byte-identical ciphertext + kek_id + updated_utc, decrypt-verified on the partner.
2. No cross-pair propagation — absent from site-b and central across 8 samples /
240s / 48 announce intervals, and clean in the reverse direction too. Isolation
controls seeded on site-b-1 and central-1 converged within their OWN pairs, so
the absence is isolation rather than a dead replicator.
3. Delete propagates within the pair — issued on site-a-2 (the receiving node, so
this also covers the reverse intra-pair direction); tombstone on site-a-1 in <1s.
4. Clean pair restart (both together) — mesh re-formed in 7s, no resurrection of the
tombstone, live secret survived and decrypted correctly on both nodes.
Zero secret-related warnings or errors fleet-wide across the whole run. Satisfies the
scadaproj#3 precondition for the OtOpcUa production topology (Akka pair-local).
Secrets:Replication:Enabled remains default-false in the product; the rig KEK is dev-only.
Rig config: SqlitePath moved onto each node's existing per-node /app/data volume. The
appsettings default resolves to /app, inside the image's writable container layer, so
the store was destroyed by any `up -d --build` recreate and was unreachable from the
host — seeding or inspecting a node meant stopping the container to docker cp the file
out. Still one private store per node; replication is still what syncs them.
Residuals in the gate doc: the live-push path is not covered (anti-entropy is what a
driver-only node can be gated on), and appsettings.json still carries a stale
"KNOWN NON-FUNCTIONAL against 0.2.0 / do not enable" comment that this gate disproves
— left uncorrected here because this run is config-only.
Claude-Session: https://claude.ai/code/session_014WNM4vjoVksyyBraTXSZE1
Ran on master @ 0b194003, post-deletion image d2a290fc5711 (provenance-guarded).
PSK matrix 7/7 incl. unset-key boot refusal; command matrix with Restart/Reconnect
driven from the real /hosts panel (#546 closed the 1A composition carry-forward by
direct evidence); exactly-once ack matrix sealed 6/6; central-kill + site-kill drills
honest (one loud WRN, +2:00 deadlines, boot-replay, no retro-ack); mid-deploy
coordinator kill: recovered coordinator fired the deadline at exactly +2:00, all
landed acks preserved, zero dupes; boundary: 3 established :4053 associations, all
pair-internal, zero site↔central; full-rig restart transport grep = 0 on all six
(one non-mesh observation: central pair reproduced the documented guard-off
cold-start split-brain, healed by single-node rejoin); MeshGrpcBoundaryTests 3/3;
terminal full suite RC=0 direct capture, 50/50 suites, 9,333/0.
I-4 cold-PSK-cache re-carried out of the program (unmeasurable on docker-dev).
No new defects.
Claude-Session: https://claude.ai/code/session_014WNM4vjoVksyyBraTXSZE1
Task 3.1: 7 ordered deletion steps + MeshPaths review fix, spec+code reviews closed.
Task 3.2: docs sweep + accuracy review, scadaproj index pushed on its main.
DoD full-suite: run 3 exit 0 (50 suites, 9326 passed); runs 1-2 each hit one
pre-existing unrelated timing flake (3 distinct tests, all untouched by the
branch, 9/9 green in isolation). Also records the exit-code-capture trap found
in run 1 and its retroactive caveat on Phase 2's check-8 wording.
Merge to master is user-owned — STOPPED at the Phase 3 DoD.
Claude-Session: https://claude.ai/code/session_014WNM4vjoVksyyBraTXSZE1
Node→central ApplyAcks over gRPC, gated live on the docker-dev rig at
4a2a3d3a. All 7 checks PASS (the 6 planned plus the mandated 4b hang
drill).
The 4b drill is the one that earns its place: with central-1 PAUSED
rather than stopped, both site-a nodes burned a full AckTimeoutSeconds
on it as DeadlineExceeded and still reached central-2 with a live
deadline — the per-endpoint FRESH deadline from 667639e6, proven on the
failure shape a hoisted deadline could not survive. Observed failover
latency ~10 s, exactly one AckTimeoutSeconds.
It also produced the run's only duplicate ack, and that turned into
evidence rather than a defect: the frozen central had in fact received
both acks and replayed them on unpause, and the coordinator reported
2/6 acks landed, not 4/6 — the idempotent absorption the sender's
at-least-once remark asserts, observed instead of argued.
Coexistence is per pair in one deployment: site-a's rows are Applied
with matching ReportApply lines on central and zero receptionist
associations, site-b's are Applied with zero ReportApply lines and a
live receptionist association. Neither pair could have used the other's
transport.
Carry-forwards recorded: the I-4 cold-PSK-cache latency is structurally
unmeasurable on docker-dev (ClusterPsks is read ahead of the store, so
InvalidateAll clears an empty cache) and moves to a store-backed rig;
a restarted central's cached channel can stay in backoff ~2 min on the
node that watched it die, absorbed by the walk with no ack lost. The
1A finding-5 wording fix gains a second confirmation — 4 of 4 admin
boots lost the PreStart race, and every ack central-1 accepted proves
the provider it said would NEVER be published was.
Docs only; the rig flip and the four Serilog Debug overrides were
reverted and are not in this diff. Rig restored to six healthy,
all-ClusterClient, sealing 6/6.
Claude-Session: https://claude.ai/code/session_014WNM4vjoVksyyBraTXSZE1
Re-ran the 1A gate on the docker-dev rig at e54883a8 after the defect-4 boot
fix. Image rebuilt (otopcua-host:dev -> 1d8452a29018); six host nodes plus
both one-shots recreated; run 1's preserved image was stale and deliberately
not reused. VERDICT: PASS.
Boot: 6/6 nodes healthy INCLUDING the central pair, 0 Hosting failed to start,
/health/ready 200 on every node, MAIN mesh 2/2.
Check 1 PASS: one DispatchDeployment, one correlation id, accepted by all six
nodes; central logs 6 accepted / 0 failed; zero ClusterClient lines on either
central node with Akka.Cluster.Tools.Client at Debug.
Check 2 PASS: deployment Sealed with all six NodeDeploymentState rows Applied
inside ~90 ms, while the four site nodes hold receptionist associations — the
only ack path they have in 1A. Mixed transport proven.
Check 4 PASS: Ack clicked on the real /alerts page against a live ACTIVATED
row; central fanned one AlarmCommand to all six over gRPC and the condition
acknowledged 0.54 s later carrying the operator identity.
Check 7 PASS: with site-b-2's GrpcPort NULLed, both central nodes warned once
and the next dispatch reached exactly five nodes; the deployment then honestly
ended TimedOut with site-b-2 alone unapplied. Row restored, 6/6 sealed after.
Check 8 PASS and no longer vacuous: 0 denials on all four interceptors, and
this time config-serve (28 KB verified artifact fetches from :4055) and
telemetry (streams open on all six, live driver health on /hosts) both carried
real traffic.
Check 3 closed by composition, with the reason now established rather than
assumed: <DriverStatusPanel has 0 usages repo-wide, no rendered page carries a
Restart/Reconnect control, and the only non-script-analysis POST is
/api/deployments. Run 1 drove the node-side leg for both commands; check 4
drives the identical AdminOperationsActor.Dispatch -> DispatchOverGrpc origin
path from the AdminUI.
Checks 5 + 6 stand from run 1 — nothing in the fix touches the interceptor.
New finding 5, recorded because it is load-bearing for the fix's design: the
PreStart resolve LOSES its race on real boots — 2 of 3 observed admin boots
logged Failed to resolve for both dependencies with the same
AuditWriterActorKey cause. The memoize-on-first-use hardening is therefore not
theoretical, and its heal was proven live by moving the coordinator singleton
onto the node that lost the race (stop central-2) and watching central-1 fan a
dispatch to all five live nodes with no dialer error. Residual is cosmetic and
carried to Phase 2: the Error text reads as permanent, and it will trip a soak
error-grep on a healthy node.
Defect 4 marked CLOSED; run 1's record kept in full underneath the combined
table. Rig restored: flip reverted, six nodes healthy on all-ClusterClient on
the committed image, three mesh ClusterClients back on each central node.
Joseph Doherty