Task 26 — the P2 milestone gate. Ran the live /run verification on an isolated
docker-dev rig (project otopcua-mqtt, ports 9210/4850-4851/14350, image built
from this branch) against the real Mosquitto TLS+auth broker and the C#
Sparkplug edge-node simulator on 10.100.0.35, and recorded the result.
The gate found three defects. Two are the same defect class the P1 gate found
twice — a hand-maintained AdminUI surface left behind by a driver-side feature —
and the third is a pre-existing cross-driver bug that only a Sparkplug flow
could surface.
1. CRITICAL, FIXED — MqttDriverForm still shipped its P1 Sparkplug PLACEHOLDER.
Switching Mode to SparkplugB rendered a "not available yet" notice and NO
Group ID field, so with Sparkplug ingest fully shipped there was still no
way to author a Sparkplug driver from the AdminUI at all. Sparkplug.GroupId
is the driver's entire subscription filter (spBv1.0/{GroupId}/#): blank ⇒
connected, Healthy, ingesting nothing. Now authors all five Sparkplug keys,
MERGES over the existing sub-object rather than replacing it (so a key a
newer driver adds inside it survives an older AdminUI), leaves it untouched
in Plain mode, and validates the group id as the topic segment it is.
Pinned by 8 MqttDriverFormModelTests cases, verified falsifiable — 8 RED
with the fix stubbed out.
2. PRE-EXISTING + CROSS-DRIVER, FIXED — every node label in the shared
DriverBrowseTree was an <a href="#">. In a Blazor Web App, blazor.web.js's
enhanced-navigation click interceptor resolves a bare "#" against
<base href="/">, so clicking ANY browse-tree node label navigated the whole
AdminUI to "/", tore down the circuit, and destroyed the hosting modal —
losing the browse session AND the tag selection. @onclick:preventDefault
does not help: it suppresses the browser's default action, not Blazor's own
interceptor. Labels are now <button type="button">. Dates to the component's
introduction (2026-05-28) and affects EVERY driver's picker; it surfaced
only now because choosing a rebirth scope is the first flow that requires
clicking a label rather than the ▶ toggle (always a button, always fine).
3. FIXED (mitigation) — the browse tree rendered exactly once, at open, and
never again, while OpenAsync returns as soon as the SUBSCRIBE is granted.
On Sparkplug that means it is almost always empty forever, because births
are never retained. Added a Refresh button that re-reads the root against
the SAME session (nothing reconnects, nothing observed is lost, the tag
selection is kept, only the armed rebirth scope is cleared).
Live gate result (all against the real broker + simulator, group OtOpcUaSim):
- driver authored end-to-end through the fixed MqttDriverForm; blob is
Mode:SparkplugB + Sparkplug.GroupId:OtOpcUaSim with enums as names
- browse: BROWSER OPEN chip, Sparkplug-only rebirth panel, tree renders
Group → EdgeNode → [Device] → Metric with the device folder Filler1 and
node-level metric leaves SIDE BY SIDE, and "Node Control/Rebirth" as ONE
leaf (the metric-name-is-not-a-topic-segment rule holding)
- rebirth: group scope names the group + the 32-node cap, Cancel published
NOTHING (simulator log unchanged); edge-node scope → "1 command published"
and the simulator logged "rebirth NCMD #1 received; republishing metadata";
group scope → "2 commands published", both edge nodes re-birthed; a device
and a metric both resolve UP to "edge node EdgeA"; the panel does NOT render
for a Plain MQTT device
- browse-commit wrote bindable tuples with no topic/address key —
{"groupId","edgeNodeId","metricName","deviceId"} for the device metric and
deviceId ABSENT (not blank) for the node-level one, datatypes inherited from
the birth (Int64 / Float)
- deployed (Sealed), and both nodes served live changing values through OPC UA
at the 2 s cadence — Good 0x00000000, no BadNodeIdUnknown
- death→STALE: stopping the simulator drove both nodes to 0x80000000 with an
empty value ~2 s later, and restarting it recovered them to Good on the new
birth (FillCount back to its birth-declared 1000, then counting)
- rediscovery fired exactly TWICE (once per authored scope: OtOpcUaSim/EdgeA
7 metrics, OtOpcUaSim/EdgeA/Filler1 3 metrics) and was then gated across
many subsequent births/rebirths — the anti-storm change gate working, and
EdgeB correctly filtered out as an unauthored scope
- tag editor: opens in SparkplugB (mode inference on reopen) with all four
descriptor fields populated; blank group → "A Sparkplug group ID is
required."; "Plant/1" rejected; "Node Control/Rebirth" ACCEPTED and saved
with the slash whole; dataType override persists and its key is ABSENT again
after selecting "(from birth certificate)"
Two things are NOT verified, and are recorded rather than claimed:
- Rediscovery is INERT in v3 and this is a platform gap, not an MQTT one:
nothing subscribes to IRediscoverable.OnRediscoveryNeeded and
DriverHostActor.HandleDiscoveredNodes hard-returns. A DBIRTH introducing a
metric does NOT change the OPC UA tree; redeploy. Verified from the driver's
log line only, exactly as far as it can be verified.
- A metric name containing "/" cannot be BROWSE-COMMITTED: the derived raw tag
Name is a RawPath segment and may not contain "/", so the spec-mandatory
"Node Control/Rebirth" is refused ("Row 3: Name must not contain '/'"). The
refusal is loud and all-or-nothing, never a silently mis-bound tag, and
Manual entry is the working path. Fixing it needs a name-sanitisation policy
with a collision answer, so it was recorded rather than guessed at.
Also left open + documented: an empty browse tree cannot arm a rebirth (the
scope comes from a tree click, and the session side already accepts a bare
{group}/{edgeNode} — only a UI path is missing); _canRebirth is captured at
browse-open; the tag editor injects the Plain-only payloadFormat key into a
Sparkplug blob (cosmetic — the factory routes on the tuple).
Suites: offline 1528 passed / 0 failed (581 Driver.Mqtt + 809 AdminUI + 138
Core.Abstractions); live 15/15 against the broker + simulator.
Docs: docs/drivers/Mqtt.md brought fully up to date for P2 (Sparkplug config
sub-object, both tag shapes, the ingest state machine, rediscovery, browse +
rebirth + Refresh, a Known-gaps table); Mqtt-Test-Fixture.md gains the simulator
and drops its "Sparkplug has no fixture" claims; infra/README.md §3 gains the
simulator row; CLAUDE.md gains the simulator endpoint facts; the tracking doc
marks MQTT/Sparkplug COMPLETE. docker-dev/docker-compose.mqtt.yml is the
isolated-rig overlay the gate ran on.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Two gaps the Task 23/24 review found, both blocking Task 26's live gate.
Gap 1 — browse-commit produced a silently dead MQTT tag. RawBrowseCommitMapper
had no `Mqtt` case, so a committed leaf fell through to the generic
`{"address": …}` key. Neither MqttTagDefinitionFactory entry point reads
`address`: the tag deployed clean and reported BadNodeIdUnknown forever, with
no signal at commit time. Predates Task 23 (Plain was affected too), but Task 23
built the Sparkplug metric tree precisely so an operator could browse and commit
a binding.
The address is a DESCRIPTOR, not a reference string, and it cannot be recovered
from the browse node id: `{group}/{node}[/{device}]::{metric}` where a metric
name legitimately contains `/` (`Node Control/Rebirth`) — the ambiguity
MetricSeparator's remarks already name. So the session STATES it, via a new
`AttributeInfo.AddressFields` seam, and the mapper reads it. Which keys are
emitted is also how the mapper learns Plain vs Sparkplug — the driver type
reaching it is just `Mqtt`, and the mode lives on the driver config. Key names
are single-sourced in the new `MqttTagConfigKeys` (producer, mapper, factory),
with the literals pinned by a test so a symmetric rename cannot silently unbind
already-persisted blobs. A leaf with no stated address is refused at commit in
words rather than committed dead.
Gap 2 — RequestRebirthAsync had no UI. Task 23 shipped it backend-only; Task 26
step 1 assumes the button, and Task 25's runbook notes the picker tree stays
empty until a birth lands, so it is the only way to fill it on demand. Added to
the /raw browse modal: scope = the clicked tree node (device/metric resolve up to
their edge node, group fans out and is refused whole past 32), an explicit
two-click confirm naming the resolved scope and its consequence, the outcome or
the refusal shown rather than swallowed. Offered only for a Sparkplug session
(new `IRebirthCapableBrowseSession.RebirthAvailable` — one session class serves
both modes, so a type test alone would draw a button that can only throw) and
only to a DriverOperator; the server-side gate remains the boundary.
Tests: MQTT 545 → 581, AdminUI 781 → 800. The round-trip tests feed the emitted
blob to the REAL factory and were falsified by mutating the emitted key name.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Unseals the Sparkplug branch of the MQTT address-picker browser that Task 10
deliberately sealed shut, and adds the first (and only) sanctioned publish on a
browse session.
- MqttBrowseSession serves Group -> EdgeNode -> [Device] -> Metric in Sparkplug
mode, decoded from observed NBIRTH/DBIRTH certificates (the only Sparkplug
messages that name their metrics). Node-level metrics hang directly under
their edge node -- no synthesised device folder, because the authored address
tuple genuinely has deviceId = null for them. Metric node ids use a '::'
separator: a metric name may contain '/', so slash-joining would make a
node-level metric indistinguishable from a device folder.
- AttributesAsync is purely birth-derived in Sparkplug mode. The plain path's
UTF-8 inference / payload-snippet machinery does not run and no payload bytes
are retained: a Sparkplug body is protobuf and would be reported as
"(N bytes, binary)" for every metric in the plant. A datatype ToDriverDataType
cannot map is reported as the Sparkplug type name, never coerced.
- RequestRebirthAsync(scope) is the one publishing member, reached only by an
explicit operator action. It routes through the counted PublishAsync
chokepoint via a private RebirthTransport adapter, so PublishCountForTest
still proves that OpenAsync/RootAsync/ExpandAsync/AttributesAsync/Observe/
DisposeAsync publish nothing -- now in both modes. A device or metric scope
resolves up to its owning edge node (NCMD is node-addressed); group scope
enumerates observed edge nodes and is refused whole past
MaxGroupRebirthNodes = 32.
- BrowserSessionService.RequestRebirthAsync gates on the same DriverOperator
policy that gates the picker, evaluated server-side where the action happens
rather than only at render time, fails closed, and Info-logs the scope and the
published count. Exposed through the new IRebirthCapableBrowseSession contract
so "does this session write?" stays a type question.
- ToBrowseOptions' Last-Will landmine re-verified: MqttDriverOptions still
carries nothing will-shaped (an NDEATH is a broker-published will and would be
invisible to PublishCountForTest), now pinned by a reflection guard.
Falsifiability: injecting a publish into RootAsync reddens the Sparkplug and
plain passivity tests; double-publishing per target reddens all four
one-NCMD-per-node assertions. 572/572 MQTT tests, 24/24 AdminUI browsing tests.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Task 24 (P2): replaces the P1 Sparkplug Validate() stub (always null) and the
"not available yet" editor placeholder with real groupId/edgeNodeId/deviceId/
metricName authoring, mirroring MqttTagDefinitionFactory.FromSparkplugTagConfig
in both directions -- required ids + optional deviceId, dataType/qos read
strictly but only when present (dataType is optional: the birth certificate
declares it). One rule is deliberately stricter than the factory: group/edge-
node/device ids reject '/', '+', '#' (topic-segment characters a decoded
incoming id can never carry, so an authored one that does can never bind);
metricName is exempt since Sparkplug's own canonical names use '/' (e.g.
"Node Control/Rebirth"). No FullName key is written -- TagConfig carries no
identity under v3, and the plan's snippet asking for one was corrected per the
brief. A SparkplugB tag now survives save/reopen because its descriptor keys
re-infer Mode on load; there was never a 'mode' key to persist.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
The P1 live gate had to insert MQTT driver config directly via SQL: DriverConfigModal and
DeviceModal both fell through to "No typed config form for driver type Mqtt" with no raw-JSON
fallback, so broker host/port/TLS/credentials were unauthorable from the AdminUI. Task 12 built
the *tag* editor; these are the driver + device surfaces.
MqttDriverForm authors the whole MqttDriverOptions connection surface (host/port/clientId,
TLS + CA pin, credentials, protocol version/clean-session/keep-alive, connect timeout, reconnect
backoffs, mode, maxPayloadBytes, and the Plain sub-object). The Sparkplug sub-object stays P2 —
a marked placeholder mirroring MqttTagConfigEditor's stub, with any existing sparkplug keys
preserved untouched.
The connection is authored on the DRIVER, not the device — unlike Modbus/S7/OpcUaClient and
contrary to what docs/drivers/Mqtt.md said. DriverDeviceConfigMerger merges a device's keys up
only when the driver has exactly ONE device, so a device-authored broker connection vanishes
silently the moment a second device is added. MqttDeviceForm is therefore informational (the
GalaxyDeviceForm shape) and round-trips DeviceConfig verbatim; it flags legacy connection keys
left on a device by a pre-form deployment, because those still win via merge-up.
Serialization goes through the ONE shared MqttJson.Options instance (Task 9's decision), never a
fifth per-form copy — pinned by an assertion that an ordinal-only reader CANNOT bind the emitted
blob, and by extending the fleet-wide DriverPageJsonConverterTests guard to resolve a form's
serializer from an explicit external-instance registry when it has no _jsonOpts field. Dropping
the converter from MqttJson.Options reddens 3 tests and leaves the symmetric round-trip green —
the Task 9 lesson, reproduced.
RawTags is stripped from the emitted blob (the deploy artifact owns it) and unknown top-level
keys survive a load->save. Validation mirrors the driver's own [Range] bounds inline, and
ToOptions() additionally clamps, so an ignored error still cannot persist a
connectTimeoutSeconds:0 driver-brick. A non-blank clientId raises the P1 live-gate warning
inline (fixed ids make redundant pair nodes evict each other while both report Healthy).
AdminUI 761/761 green (was 730), TWAE-clean; Driver.Mqtt 266/266 green.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Task 14 — the P1 milestone gate. Ran the live /run verification on a docker-dev
rig against the real Mosquitto TLS+auth fixture, and recorded the result.
The gate did its job: it found MQTT was unauthorable in production.
RawDriverTypeDialog's driver-type list is a hand-maintained array that nobody
added MQTT to, so the /raw "New driver" picker never offered it — with every
other layer (contracts, driver, browser, factory, host registration, typed tag
editor) complete and 266 green unit tests. Nothing tied that array to
DriverTypeNames and AdminUI has no bUnit, so no offline test could see it.
Fixed, plus RawDriverTypeDialogParityTests: a reflection parity guard that
fails for ANY DriverTypeNames entry missing from the picker. Verified
falsifiable — RED with "…cannot be authored from the /raw New-driver picker,
so they are unreachable in production: Mqtt" before the one-line fix.
A second gap is left OPEN and documented, not fixed (no task ever covered it):
MqttDriverForm / MqttDeviceForm do not exist, and DriverConfigModal/DeviceModal
have no raw-JSON fallback — so an operator can create an MQTT driver but cannot
author its broker endpoint or credentials. P1 is not operator-complete until
those two razor forms land. The gate authored config via SQL to get past it.
Live gate result (isolated 2-node MAIN stack, image built from this branch,
fixture at 10.100.0.35:8883, AllowUntrustedServerCertificate rather than
pinning the CA into the rig):
- typed MQTT tag editor renders and round-trips; Json shows the JSON-path
field, Raw/Scalar hide it; wildcard topic a/+/c blocked at save; data-type
list offers Float64 and no Double; qos/retainSeed absent on "(driver
default)" and qos:2 written as a number; isHistorized/historianTagname
survive a topic-only edit; SparkplugB renders the P2 placeholder and
switching back keeps Plain values; jsonPath is guidance not a gate in all
three shapes ($ seeded only for a brand-new blank tag, blank stays absent,
clearing saves fine)
- deployed, and both editor-written blobs resolved and served changing live
values through OPC UA (22.8 / 1629, StatusCode Good) — no BadNodeIdUnknown
- the bespoke #-observation browser drove a real broker session and rendered
the fixture's actual topic tree (line1/{counter,state,temperature},
noise/chatter, retained/seed); a Modbus device's picker still correctly
reports "Browsing unavailable"
Also found live and documented (not a code change): both nodes of a redundant
pair run the driver, so a FIXED clientId makes them evict each other forever —
the broker logs "already connected, closing old connection" and both nodes
reconnect every ~2s while still reporting Healthy. Unset (the default) is
correct; confirmed 0 reconnects/60s on both nodes after removing it.
Suites: offline 1134 passed / 0 failed (266 Driver.Mqtt + 730 AdminUI incl. the
3 new guards + 138 Core.Abstractions); live 7/7 against the fixture.
Docs: new docs/drivers/Mqtt.md + Mqtt-Test-Fixture.md (the per-driver + fixture
convention every other driver follows), MQTT rows in docs/drivers/README.md,
TestConnectProbes.md, root README.md, and infra/README.md §3.
CLAUDE.md drift corrected while adding the MQTT endpoints:
- the "every fixture carries project: lmxopcua" claim was false — only the
MQTT fixture does; the filter returns one stack, not the fleet
- lmxopcua-fix.ps1 is Windows-VM-only; on macOS drive the host over ssh/rsync
(and the docker-dev rig itself runs locally under OrbStack)
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Review follow-up. Dropped the hard Validate rejection of a blank jsonPath under
a Json payload: the runtime defaults it to the document root, which is the real
"publisher puts a bare JSON scalar on the topic" case, so rejecting it made the
editor refuse what the driver accepts — inverting "authoring surface accepts <=>
publish accepts" — and blocked whole CSV-import batches, since this validator
also gates RawManualTagEntryModal's review grid.
Replaced with guidance: an UNAUTHORED tag (no topic AND no jsonPath) seeds the
field with "$", so a fresh tag emits an explicit "jsonPath":"$" while an
existing path-less tag keeps the key ABSENT and the driver defaults it. The
wildcard-topic rule stays strict — a wildcard has no legitimate single-Tag
interpretation.
Also: omit a blank "topic" rather than writing "topic":"", and record why the
_lastConfigJson guard diverges from the Modbus template — the landmine is a
DERIVED, NON-PERSISTED UI FIELD INFERRED FROM BLOB CONTENT (Mode), which Task
24's Sparkplug field group will be tempted to add more of. Named the host-side
dependency (RawTagModal only mutates through this callback) that keeps the
staleness trade benign today.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Thin typed model over a preserved JsonObject key bag, mirroring the sibling
<Driver>TagConfigModel template. Key names and strictness rules mirror
MqttTagDefinitionFactory rather than inventing an editor-side schema; enums
round-trip as NAMES (the factory reads them strictly by name).
No FullName key: under v3 a tag is identified by its RawPath, which the factory
keys the definition's Name off — a composed identity key here would be dead
weight nothing reads.
Mode is a UI-only sub-shape selector inferred from the blob (any Sparkplug
descriptor key present) and never serialised — the driver takes Plain vs
SparkplugB from its DRIVER config. Only Plain Validate() is implemented; the
Sparkplug branch is a stub Task 24 fills, and its keys are preserved untouched.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
ClusterPrimaryHealthCheck was written three days' worth of debugging ago because
the shared ActiveNodeHealthCheck selected by RoleLeader and reported Healthy for
any node lacking the role. Health 0.3.0 fixes the shared check — oldest Up
member, role-preference scoping, Unhealthy when the node owns no active work —
so the private copy is now duplication rather than a workaround, and it is
deleted.
RolePreference [admin, driver] reproduces exactly what it did: a fused central
node answers for admin (where the singletons and the AdminUI are pinned), a
driver-only site node answers for driver.
SelectOldestUpMemberOfRole now delegates to the shared ClusterActiveNode instead
of re-implementing the age ordering. This is the point of the change rather than
a tidy-up: the redundancy snapshot drives the OPC UA ServiceLevel 250/240 split
while the health tier drives Traefik's admin routing, so two copies of "oldest Up
member of a role" would be two chances to advertise one node as authoritative
while gating the data plane on another. ControlPlane takes a ZB.MOM.WW.Health.Akka
reference for it; the layering trade-off is recorded at the PackageReference.
Behaviour note: a node carrying neither admin nor driver now answers 503 on the
active tier instead of 200. That case is reachable — RoleParser admits dev-only
and cluster-role-only nodes — and 503 is correct, since such a node owns no
active work and must not be in an active-tier pool. No docker-dev node is
affected: every rig node is admin+driver or driver-only.
Tests replaced in kind. The rule itself is now pinned in the library against a
real two-node cluster; what stays OtOpcUa's own decision is the wiring, so
ActiveTierRegistrationTests pins which check is on the active tag, that it is
registered on driver-only nodes too, and that it is not on the ready tier.
Verified: Host.Tests 25/25, ControlPlane redundancy 15/15.
Closes the remaining half of #494, and corrects the half fixed in 8dd9da7d.
The shared ActiveNodeHealthCheck(role: "admin") answered the wrong question twice
over. It returns Healthy for any node LACKING the role, so all four driver-only
site nodes called themselves active and no consumer could find the Primary of a
site Cluster. And it selects by RoleLeader - the lowest-ADDRESSED member - which
is not where Akka places singletons.
Replaced with ClusterPrimaryHealthCheck ("cluster-primary"). One rule: this node
is active iff it is the OLDEST Up member carrying its own active role, where the
active role is admin when the node has it and driver otherwise. That serves both
consumers correctly - a fused admin node answers for admin, so Traefik pins the
AdminUI to the node hosting the singletons; a driver-only site node answers for
driver, which is exactly SelectDriverPrimary, the same election behind
IsDriverPrimary and the OPC UA ServiceLevel 250/240 split. Per-mesh scoping is
free after Phase 6: ClusterState.Members already contains only this node's own
Cluster.
SelectDriverPrimary is generalised to SelectOldestUpMemberOfRole so the
age-ordering rule has one implementation. Two copies of "oldest Up member of a
role" would be two chances to silently disagree about who is in charge, and the
tier and the redundancy snapshot must never disagree.
THE ROLE-LEADER HALF MATTERED IN PRACTICE, not just in theory. On the rebuilt rig
the two orderings diverge right now:
akka leader (lowest address) = central-1
oldest admin member = central-2 <- hosts the singletons
8dd9da7d made the tier a real 503 but still selected by RoleLeader, so it would
have pinned Traefik to central-1 - the node NOT hosting the work. With this change
Traefik correctly routes to central-2.
Live-verified, exactly one 200 per mesh:
MAIN central-2 200 central-1 503 traefik: central-2 UP, central-1 DOWN
SITE-A site-a-1 200 site-a-2 503
SITE-B site-b-1 200 site-b-2 503
Tests: role-selection matrix and the startup-safe Degraded path in Host.Tests
(26 pass); parity between the tier's selector and the redundancy snapshot's, plus
role scoping, added to RedundancyPrimaryElectionTests, which forms a real two-node
cluster deliberately built so the oldest member is not the lowest-addressed one
(5 pass).
Phase 6 wired the re-homed redundancy singleton in Program.cs as
`WithOtOpcUaClusterRedundancySingleton(sp.GetRequiredService<IClusterRoleInfo>())`
inside the AddAkka configurator lambda. That lambda runs WHILE the ActorSystem is
being built, but ClusterRoleInfo depends on the ActorSystem (it is a live
Cluster.State view) — so resolving it there recurses into ActorSystem construction
and every node dies at boot with StackOverflowException. It compiles cleanly (a
runtime DI cycle) and RedundancyStateSingletonRehomeTests passed a hand-built
FakeClusterRoleInfo straight into the extension, mocking around the composition-root
line that overflows. The docker-dev live gate caught it on first boot.
Derive the singleton's cluster-role scope from AkkaClusterOptions (pure config, no
ActorSystem) instead: BuildClusterRedundancySingletonOptions / the extension now take
AkkaClusterOptions and compute Role = Roles.FirstOrDefault(IsClusterRole) ?? driver —
identical to ClusterRoleInfo.ClusterRole, which derives from the same config. Program.cs
passes IOptions<AkkaClusterOptions>.Value, which has no ActorSystem dependency.
Live-verified: rebuilt image boots with zero stack-overflow lines, cluster singletons
form per mesh. 5/5 rehome tests pass; Host builds clean.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Task 3 review: scoping only the rows/enabled DB queries by ClusterId left
the observed set (from Cluster.State.Members) filtered by the driver role
alone. In the mixed-mesh window — DB scoping live via a cluster role, but
the physical Akka mesh not yet split so central still sees foreign members
via gossip — a foreign-cluster driver member stays in observed while its row
is filtered out, hitting the RunningNodeHasNoRow branch, which logs at ERROR.
That traded the Warning-level false positive for a worse Error-level one.
Filter observed by the node's own cluster ROLE too. Extracted the projection
into a static FilterOwnClusterDriverMembers(members, ownClusterRole) so the
membership filter is unit-testable without seeding a live cluster (the actor
harness joins as admin only, so observed is always empty). Sourced at
registration from IClusterRoleInfo.ClusterRole alongside ClusterId. Null role
(legacy, no cluster role) keeps every driver member — reconcile-all, unchanged.
Also clarified the ownClusterId doc wording (non-null-and-non-empty).
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Review follow-ups on the redundancy re-home (Phase 6, Task 2):
- Program.cs: the hasDriver comment claimed IClusterRoleInfo "throws at host
start if this driver node carries no cluster role" — false since the fallback
landed. Rewrite it to describe the cluster-{ClusterId} scope with the driver-role
fallback for legacy single-mesh / not-yet-migrated nodes.
- ClusterRoleInfo.cs: the SubscriberActor comment described RedundancyStateActor as
the "admin-role singleton" — stale. Note it is now a cluster-{ClusterId}-scoped
singleton spawned on every driver node (LeaderChanged stays a no-op here).
- Add a host/registry-level test for the driver-role FALLBACK path (roles ["driver"],
no cluster role) closing the coverage asymmetry — it was proven only in the pure
helper. Asserts the node boots and registers RedundancyStateActorKey, i.e. the boot
the earlier throw-based version would have aborted.
The boot helper now supplies an in-memory IDbContextFactory and a fake IClusterRoleInfo:
Akka.Hosting invokes singleton props factories eagerly at StartAsync, and the admin
ClusterNodeAddressReconciler factory reads both (the latter for its per-cluster reconcile
scope, added by a sibling Phase 6 task) — without them the admin boot NREs.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
After the fleet splits into one Akka mesh per application Cluster, a
ClusterClientReceptionist serves only its own mesh — so the single
fleet-wide client's SendToAll reached only the mesh whose receptionist
answered, leaving every other cluster silently on its old configuration.
Central now holds one ClusterClient per ClusterId and fans SendToAll
across all of them. LoadContactsFromDb selects ClusterId and groups the
receptionist contacts per cluster (keeping the per-row TryParse guard and
the enabled/non-maintenance filter); ContactsLoaded carries
ContactsByCluster; HandleContactsLoaded diffs per cluster (rebuild changed,
stop+drop vanished, warn-don't-create on empty); RebuildClient builds a
per-cluster client under a clusterId-derived actor name. IMeshClusterClientFactory.Create
gained a clusterId parameter so the per-cluster clients get distinct,
diagnosable names. ApplyAck handling and the DPS branch are unchanged; the
deploy path stays payload-free.
Tests: focused unit coverage of the grouping + fan-out (one-client-per-cluster,
fan-out SendToAll to every cluster client) via the recording factory double,
plus the real two-mesh boundary test extended to central + two separate site
meshes proving one dispatch reaches both and a client scoped to one cluster
never crosses into another. Red-before-green verified: crippling the fan-out
to a single client fails the reaches-both-meshes assertion.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 6 splits the fleet into one mesh per Cluster, so an admin (central)
node's Cluster.State.Members shows only its own pair — it no longer sees
site members via gossip. The ClusterNodeAddressReconciler singleton compared
live membership against EVERY ClusterNode row fleet-wide, so post-split every
foreign-cluster row would log EnabledRowNotInCluster forever.
Scope the actor's two DB queries to rows whose ClusterId matches the admin
node's own (IClusterRoleInfo.ClusterId, sourced at registration). A legacy
admin node with no cluster role (null ClusterId) still reconciles the whole
fleet — it genuinely sees every member via gossip. The pure Reconcile
function and AddressMismatchKind semantics are unchanged; scoping lives
entirely in the actor's queries.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 6 splits the single fleet-wide Akka mesh into one 2-node mesh per
application Cluster. RedundancyStateActor was an admin-scoped cluster singleton
that elected ONE fleet-wide driver Primary — but after the split a driver-only
site pair has no admin node in its mesh to host it, so its Primary would never
be elected.
Re-home it: remove the redundancy WithSingleton block from the admin
WithOtOpcUaControlPlaneSingletons set, and add WithOtOpcUaClusterRedundancySingleton,
a singleton spawned from Program.cs's hasDriver branch on EVERY driver node (the
fused central included). It scopes to the node's own cluster-{ClusterId} role when
present, and falls back to the fixed `driver` role otherwise. Each mesh then has
exactly one, electing its own pair-local Primary and publishing redundancy-state
on its own mesh's DistributedPubSub. The election logic (oldest Up driver member)
and the DPS publish are unchanged — they become pair-local after the split.
The driver-role fallback (Decision 2) deliberately does NOT throw when a node has
no cluster role: that is the pre-Phase-6 fleet-wide behavior on a legacy single
mesh, so legacy/harness (TwoNodeClusterHarness: admin,driver) and not-yet-migrated
deployments keep booting unchanged. On a genuinely split 2-node mesh the `driver`
role is already pair-local (the mesh IS the pair), so the fallback is correct in
both worlds. Cluster-scope, when present, additionally survives an accidental
two-mesh merge by keeping one singleton per cluster.
The role scope + driver-role fallback are pinned by unit tests against the
extracted BuildClusterRedundancySingletonOptions helper (the BuildDowningHocon
pattern); admin-removal + driver-registration are pinned behaviorally against a
booted node's ActorRegistry.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 6 Task 1. IClusterRoleInfo gains ClusterRole/ClusterId, derived from
the node's configured AkkaClusterOptions.Roles at construction time (not
live Cluster.State) so the identity is available before the cluster forms.
First cluster-scoped role wins when more than one is configured, logged as
a Warning. Updates the FakeClusterRoleInfo test double in
ServiceCollectionExtensionsTests to satisfy the new interface members.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Closes two test-completeness gaps flagged in review of the seam-tap commit — the
ScriptedAlarmHostActor and VirtualTagActor hub emits had no test proving the hub
actually receives them.
- ScriptedAlarm_activation_emits_one_Alarm_item_with_matching_payload: spawns the
host with a capturing fake hub, drives a real engine Inactive->Active transition
through the existing ScriptedAlarms harness, and asserts exactly one
TelemetryItem.Alarm carrying the Activated transition.
- VirtualTag_script_log_emits_one_Script_item_with_same_payload: passes a fake hub
via Props + a publisherFactory, drives an evaluator failure, and asserts a
TelemetryItem.Script carrying the SAME ScriptLogEntry instance handed to the
publisher.
To make the VirtualTagActor tap observable, its hub emit moved to the TOP of
PublishLog, before the test-only publisherFactory early-return branch, so the emit
fires on BOTH the production DPS path and the factory path. Production behavior is
unchanged: production passes publisherFactory: null, so it still reaches the DPS
Tell; the emit just moved a couple lines earlier (both are fire-and-forget).
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Each of the four telemetry publish seams now also emits its DPS payload into the
node-local ITelemetryLocalHub (Phase 5), leaving the existing DistributedPubSub
publish untouched:
- driver-health AkkaDriverHealthPublisher -> TelemetryItem.Health
- resilience DriverResilienceStatusPublisherService -> TelemetryItem.Resilience
- alerts ScriptedAlarmHostActor + DriverHostActor (native) -> TelemetryItem.Alarm
- script-logs VirtualTagActor + DpsScriptLogPublisher -> TelemetryItem.Script
DI services take the hub as a required ctor param; actors take an optional
nullable hub threaded through their Props and the real spawn sites
(DriverHostActor <- ServiceCollectionExtensions; VirtualTagActor/ScriptedAlarmHostActor
<- DriverHostActor), so existing test Props constructions keep compiling and simply
do not emit. The hub is a no-op until a gRPC client subscribes, so the emit is safe
on every node.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Scripted-alarm condition state lives in LocalDb (Phase 4); the ConfigDb-backed
Ef store + ScriptedAlarmState table are now dead. Removes the test-harness Ef
fallback (a DB-backed node with no store now skips the alarm host), deletes the
store + entity + model config, and drops the table via migration.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
A driver-only node has a null dbFactory but a wired applier; the old combined
guard routed it to the raw-sink fallback that wipes the address space and never
re-materialises. Split the guard so in-hand artifact bytes drive the real
diff-and-apply; a missing artifact abandons the rebuild (keep last-known-good,
#485) instead of wiping.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Nullable ConfigDb factory; UpsertNodeDeploymentState no-ops when absent
(central persists acks from the ApplyAck); scripted-alarm condition state
served from the LocalDb store instead of the ConfigDb-backed Ef store.
Combines Phase-4 Tasks 4 + 6. Removes the interim dbFactory! cast.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
The Phase 3 live gate deadlocked every FetchAndCache deploy: DeploymentArtifactGrpcService
gated on Status==Sealed, but central seals a deployment only AFTER every node acks Applied,
and a FetchAndCache node cannot ack until it has fetched the artifact — seal-needs-ack ->
ack-needs-fetch -> fetch-needs-sealed. The node's fetch reached central and passed the
shared-key interceptor, then got a clean NotFound; the #485 apply-failure path correctly
kept last-known-good, but the deploy could never seal.
Direct mode reads the same ArtifactBlob from SQL while the row is still AwaitingApplyAcks,
so the serve path must too. Drop the Sealed gate: serve any deployment whose blob is
non-empty (unknown id / empty blob still collapse to NotFound — existence-hiding + #485).
Access is already gated by the interceptor, so there is no reason to hide a non-sealed
deployment a node is legitimately applying. The Task 2 'non-sealed -> NotFound' test flips
to 'non-sealed with a blob is served'.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 3 Task 8. Stands a real h2c-only Kestrel listener serving the real
DeploymentArtifactGrpcService behind the real ConfigServeAuthInterceptor (in-memory
config DB seeded with one sealed deployment), and drives the real
GrpcDeploymentArtifactFetcher across it — proving a stream crosses the real boundary,
not just that the fetcher's logic is right (Task 4's fakes). Asserts: (1) right key +
>256 KB blob reassembles byte-equal and verifies against RevisionHash; (2) FALSIFIABILITY
CONTROL — wrong key returns null while the right key succeeds on the SAME server (so the
null is the auth gate, not a dead server); (3) unknown id -> null (NotFound-hidden);
(4) [dead-port, real-port] failover still returns the bytes. h2c GrpcChannel over http://
works out of the box in .NET 10. Tagged [Trait Category=ArtifactBoundary].
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 3 Task 7. Regression net pinning the #485 negative on every FetchAndCache seam —
nothing rebuilds, nothing is torn down on a failure: (1) a null fetch mid-steady-state
(all endpoints down / NotFound / SHA mismatch — indistinguishable to the actor) keeps
the served address space and revision (no RebuildAddressSpace for the failed revision);
(2) a zero-length blob handed to the actor fails via ReconcileDriversFromBlob's own #485
guard, independent of the fetcher's null contract; (3) a corrupt cache at boot (surfaced
as GetCurrentUnkeyedAsync == null) lands Steady-no-revision and applies NOTHING — never
a partial address space. No new mechanism; all green as a net. Removing the #485 empty
guard reddens the zero-length test.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 3 Task 6. In FetchAndCache mode Bootstrap() branches to BootstrapFromCache(),
which restores served state from the LocalDb pointer (GetCurrentUnkeyedAsync → set
_currentRevision → ApplyCachedArtifact from the cached bytes) with NO central-SQL read.
Distinct from the Direct-mode TryBootFromCache fallback: reading the cache is NORMAL
operation here so _isRunningFromCache stays false (the node can still fetch new
deploys). An empty OR unreadable cache lands Steady-with-no-revision (the first dispatch
fetches), never Stale — Stale means 'central SQL down, retry it', and FetchAndCache has
no config SQL read to recover. TryRecoverFromStale gains a defensive FetchAndCache guard
(it is unreachable in that mode, since the retry-db timer only starts in Stale).
Proven with a ThrowingDbFactory in every test: reaching Steady + applying a dispatch
proves the boot never touched central SQL. Sabotaging the mode branch (fall through to
the SQL read) reddens all three — test A flips RunningFromCache; B/C enter Stale, which
ignores the dispatch.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 3 Task 5. Under ConfigSource:Mode=FetchAndCache, a DispatchDeployment whose
revision differs from _currentRevision kicks off a gRPC artifact fetch that PipeTo's
its result back to Self as FetchedForApply (never a blocking await in a receive — that
would freeze the mailbox for the fetch deadline). HandleFetchedForApply then applies
synchronously on the actor thread, mirroring ApplyAndAck: null bytes = apply FAILURE
(keep last-known-good, do not advance the revision, ack Failed — #485); non-null bytes
reconcile drivers, rebuild the address space and push subscriptions FROM the bytes in
hand (OpcUaPublishActor never reads central SQL), then cache them. A faulted fetch task
maps to null bytes so it can't escape as an unhandled message.
Extracted ReconcileDriversFromBlob from ReconcileDrivers so Direct-read, FetchAndCache
and boot-from-cache share one reconcile body (the #485 empty guard moves into it);
ApplyCachedArtifact now reuses it instead of duplicating the spawn-plan logic.
DI: DriverHostActor.Props gains fetchAndCacheMode + artifactFetcher (LAST, per the
positional-forwarding warning); Runtime resolves them from IOptions<ConfigSourceOptions>
(fetcher only in FetchAndCache mode); Host registers GrpcDeploymentArtifactFetcher under
hasDriver.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 3 Task 4. GrpcDeploymentArtifactFetcher streams the artifact from the first
configured central endpoint that answers, reassembles the chunks, and verifies
SHA-256(bytes) lowercase-hex == the dispatched RevisionHash (byte-identical to
ConfigComposer's Convert.ToHexStringLower(SHA256.HashData(blob))). Every per-endpoint
problem — RpcException, zero-length stream, or hash mismatch — is logged and the next
endpoint tried; if none yield verified bytes it returns null (apply failure, keep
last-known-good — the #485 contract), never throwing into the actor loop. A
Func<endpoint, client> seam keeps the gRPC boundary out of the unit tests (that is
Task 8's job); the real path caches one h2c GrpcChannel per endpoint.
Files live under the .DeploymentCache namespace (folder Deployment/) to avoid shadowing
the Configuration.Entities.Deployment type in the test assembly — same convention as
LocalDbDeploymentArtifactCache.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 3 Task 3. Central serves the deployment artifact over a dedicated h2c
Kestrel listener (ConfigServe:GrpcListenPort), gated by the ConfigServeAuthInterceptor
(shared bearer key, FixedTimeEquals, fail-closed, path-scoped to
/deployment_artifact.v1.DeploymentArtifactService/). The listener block is merged
with the LocalDb-sync listener so a fused admin+driver node re-applies its existing
HTTP surface exactly ONCE and adds both dedicated h2c ports on top — double-binding
would throw 'address already in use'; zero re-binding would silently unbind the
AdminUI behind Traefik. AddGrpc now registers under hasDriver || hasAdmin with both
path-scoped interceptors sharing one pipeline.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Streams a sealed deployment's ArtifactBlob to a fetching driver node in
128-KiB chunks (matching the LocalDb cache chunk size). Unknown id, a
non-sealed deployment, and a zero-length blob all collapse to one
indistinguishable NotFound — existence-hiding plus the #485 serve-side guard
(never stream empty bytes as a valid empty config). The impl is named
DeploymentArtifactGrpcService to avoid colliding with the generated
DeploymentArtifactService container type.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Two genuinely separate single-member clusters sharing the ActorSystem name and
joined only by ClusterClient, so the mesh boundary is exercised over the wire
rather than through an in-process relay: a relay cannot fail for a missing
receptionist extension, an unregistered service, a malformed contact path or a
serialization break, which are the Phase 2 failure modes.
Covers both legs plus the control:
* central SendToAll -> node comm actor -> node EventStream
* node ApplyAck -> central comm actor, through the node's own client
* a node without the receptionist extension receives nothing
Sabotage-verified. Suppressing the node-side EventStream publish and the
outbound ack Tell reddens one leg each. Restoring the receptionist and the
service registration on the control node reddens the control, which is what
proves the control's send is live rather than silently misaddressed.
The control removes the extension AND the RegisterService call, since resolving
the receptionist to register would materialise the extension whose absence is
the point; it falsifies "delivery happens without a receptionist boundary", not
the extension line alone. Recorded in the test.
Also corrects MeshCommActorPathTests: with one node per side, per-node and
singleton registration are indistinguishable here, so the property the dropped
Task 6 assertion covered belongs to live-gate step 8, not to this class.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 2 Task 7. The five central->node publish sites -- ConfigPublishCoordinator's
deploy dispatch, and AdminOperationsActor's restart, reconnect, acknowledge and
shelve -- now hand a MeshCommand to the transport router instead of telling the
DistributedPubSub mediator directly. Neither actor knows which transport is in
force any more.
_meshRouter is NULLABLE with a mediator fallback, and that is deliberate: about
a dozen existing coordinator and admin-operations tests construct these actors
without a router, and they must keep exercising the real DPS path rather than a
silently-disabled one. The fallback is not a permanent seam -- Phase 6 deletes
it with the DPS branch and the single mesh.
Wiring note: the comm actor's WithActors block moved ABOVE the singleton
registrations so both singletons can resolve it with registry.Get, which is the
ordering-by-registration idiom AdminOperationsActor already uses for the
coordinator. An earlier version used ActorSelection.ResolveOne().GetAwaiter()
.GetResult() inside the props factory -- that blocks inside actor construction
and races the very spawn it waits for.
Sabotage-verified by inverting the branch so the mediator always wins: all six
new tests go red, including the fallback test (which proves it is asserting the
mediator path rather than passing by accident).
Suites after: ControlPlane 118/118, Runtime 450/450, Host.IntegrationTests
196/202 -- sole failure AbCip_Green_AgainstSim, the fixture baseline that fails
on master too.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 2 Task 6. Both comm actors are now spawned and receptionist-registered:
CentralCommunicationActor from WithOtOpcUaControlPlaneSingletons, and
NodeCommunicationActor from WithOtOpcUaRuntimeActors (before DriverHostActor,
because it is the host's ackRouter under ClusterClient mode; it has no
dependency of its own on the host, since inbound commands travel via the node's
EventStream).
Both are plain WithActors registrations, NOT singletons -- a driver node's
ClusterClient rotates across contact points and must find a live comm actor at
whichever node answers. Both register with the receptionist in BOTH transport
modes: an idle registration costs nothing, whereas registering only under
ClusterClient would mean flipping the flag on a running fleet needs a restart
before anything can be reached, turning a config change back into a deployment.
The coordinator ref is resolved lazily per message (Func<IActorRef?> +
registry.TryGet) so the comm actor never depends on the order in which
Akka.Hosting materialises the singleton proxy relative to this block.
Adds MeshCommActorPathTests, which boots the real two-node host and Identify-
probes both paths. These strings are the wire contract and nothing else asserts
they agree: a rename would compile, pass every unit test, deploy, and deliver
nothing, because a ClusterClient send to an unregistered path is dropped
silently. Sabotage-verified with an actual rename.
DELIBERATELY NOT TESTED, with the reason recorded in the file: that the actors
are per-node rather than singletons. Two discriminators were tried and both were
invalid -- "the path resolves on both nodes" passes under either shape because a
ClusterSingletonManager is also created on every node in the role, and "no
/singleton child" is null even for the KNOWN singleton /user/config-publish, so
Akka.Hosting does not lay them out that way. A sabotage re-registering the actor
via WithSingleton left both green, which is what exposed them. Rather than ship
an assertion that cannot fail, the property is left to the Task 8 boundary test,
where a send only arrives if the target really is registered.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Phase 2 Task 5. DriverHostActor and ScriptedAlarmHostActor now subscribe to the
node-local EventStream alongside their existing DPS topic subscriptions, so the
transport flag lives ENTIRELY on the publishing side -- exactly one of the two
ever publishes, and the switch can be flipped or reverted without touching a
single subscriber.
Note the alarm case is not symmetric: the DPS subscribe there still carries the
node-LOCAL publisher (OtOpcUaServerHostedService's OPC UA Part 9 method calls),
which never crosses the boundary and stays on DPS in both modes. Only the
central AdminUI leg moves.
Renamed DriverHostActor._coordinatorOverride to _ackRouter. Under ClusterClient
mode that ref is the NodeCommunicationActor, emphatically not a coordinator,
and the old name would send the next reader looking for one.
Adds two wiring tests, because the unit tests either side of this seam BOTH pass
with the subscription deleted -- the comm actor's test asserts a probe receives
the message, and the host tests drive the actor by direct Tell. Neither notices
a missing PreStart subscribe. Sabotage-verified: deleting either subscription
reddens exactly its own wiring test and nothing else.
Three things went wrong while writing those tests, all worth keeping:
- The first version raced. ActorOf returns before PreStart runs, and an
EventStream.Publish with no subscriber is dropped silently -- no buffering, no
dead letter. Fixed with a warm-up whose ack proves PreStart completed; the
comment says why it is not ceremony.
- The alarm version was VACUOUS: it told the host directly INSIDE the assertion
window alongside the relay, so the direct Tell alone produced the log the
assertion waited for. Split into warm-up (outside) and relay (inside).
- The alarm test needs its own ActorSystem: the shared harness pins
loglevel = WARNING and the only signal an unowned alarm gives is a Debug line.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Joseph Doherty