The default decider restarts a faulting child immediately and without limit, and
the hook makes every restart do ShutdownAsync+InitializeAsync+SubscribeAsync, so a
driver faulting on every value became a reconnect storm. Per-driver last-reprovision
stamp + DefaultReprovisionBackoff (10s, matched to DefaultReconnectInterval);
a held request is re-delivered by a single per-driver timer carrying the original
incarnation ref, so the F2 identity check still applies.
Seen RED with the backoff disabled: InitializeCount == 3 (second applied
immediately) where the test requires 2-then-3.
Also: the _children.Keys union-arm test (seeder gained taglessDrivers) and the
_desiredByDriver xmldoc note that it lingers on last-known-good after a failed apply.
A ReprovisionRequested in flight from an old incarnation could reach the NEW child
after HandleRestartDriver swapped it in. In Connecting that child DOES handle
InitializeRequested, so it started a second concurrent InitializeAsync, racing the
#527 dispose-before-overwrite.
Seen RED: InitializeCount == 2 on a child held in Connecting by a gated init
(an already-connected child swallows InitializeRequested and passes vacuously).
The alarm subscribe gated on 'a handle exists', which is correct for an idempotent
re-push and wrong for a deploy that widens/narrows the set. Combined with the
spawn-time desired set, a respawning deploy could latch the PREVIOUS revision's
alarm refs for the life of the connection; sourceNodeIds-filtering drivers
(OpcUaClient/AbCip/Focas) would never deliver the new alarm.
Seen RED: LastSubscribedAlarmRefs stayed [alarmX] after a deploy widening to
[alarmX, alarmY].
Crash seam: the planned null-Snapshot OnDataChange (no SupervisorStrategy fallback needed).
Seen RED with T6 reverted: InitializeCount 1, not 2 — with a preceding ShutdownCount>=1
assertion proving the restart really happened, so the red is not ambiguous with 'never crashed'.
Seen RED against a T3-reverted build: all 4 failed, Created[1].SubscribeCount == 0
on the headline case. 20x isolation run green after the fix was restored.
Adversarial review of 9f68f272 (classifier: non-alarm raw-tag deltas became
surgical) found stale UNS inheritance after a surgical raw edit. Editing a raw
tag's IsHistorized / HistorianTagname / IsArray / ArrayLength left every
referencing UNS node STALE: none of those fields were on UnsReferenceVariable,
so no coincident UnsReferenceDelta fired; the applier updated the raw node in
place; and EnsureVariable early-returns on an existing node. Consequences:
HistoryRead through the UNS NodeId served a different series than through the
raw NodeId (breaking the documented dual-registration invariant), a
de-historized tag kept serving history via UNS, and a scalar->array flip left
the UNS mirror scalar AND writable (bypassing the array read-only write gate).
UnsReferenceVariable now carries Backing{IsHistorized,HistorianTagname,IsArray,
ArrayLength}, copied from the backing tag's TagConfigIntent at compose time, so
record equality makes the reference Changed in the SAME plan. Both deltas are
non-rebuild kinds, so the existing machinery repairs the mirror surgically:
remove-previous in Apply, recreate in MaterialiseUnsReferences. That pass now
reads the four values from the RECORD (the single truth the planner diffed)
instead of a rawByPath back-lookup, preserving historian resolution exactly
(blank override => BackingRawPath, which is the backing tag's NodeId).
The artifact-decode seam needs no change: DeploymentArtifact.BuildRawAndUnsSubtrees
drives AddressSpaceComposer.Compose verbatim, which is what guarantees byte-parity.
The classifier needs no change either (a UnsReferenceDelta never forces Rebuild);
its comment now records why.
Also from the same review: a three-place-contract comment on RawTagPlan's
hand-written Equals naming its two downstream consumers, plus classifier pins for
the .Any(!eligible) quantifier, an ArrayLength-only delta, and an
alarm-content-only delta.
ConfigComposer gzips the serialised snapshot (CompressionLevel.Fastest)
immediately after SerializeToUtf8Bytes and BEFORE taking the SHA-256, so
the compressed bytes ARE the artifact everywhere: Deployment.ArtifactBlob,
the LocalDb cache chunks, the gRPC fetch stream, and the RevisionHash a
FetchAndCache node verifies against. The invariant the deploy path rests on
— RevisionHash == SHA-256(stored/wire bytes) — is preserved, which is what
leaves the fetcher, the cache and the serve path untouched: nothing between
the composer and DeploymentArtifact.Open ever looks inside the blob.
Read side: Open() sniffs the gzip magic 1F 8B (valid JSON can never start
with 0x1F, so no framing of our own is needed) and inflates through a size
guard — MaxDecompressedBytes, 512 MiB, internal + mutable so the guard can
be exercised without inflating half a gigabyte per test run. Pre-upgrade
uncompressed blobs take the identical path they always did.
Every decompression failure — corrupt stream, truncated stream, over-limit
inflation — is reported as JsonException, the same exception family the
malformed-JSON path already owns. So the lenient static wrappers still
degrade to empty results and the apply paths still fail the apply and keep
last-known-good (#485); no caller's decision changes because the reason the
bytes are unusable changed.
gzip output is not guaranteed byte-stable across .NET releases. Harmless
here: the hash is computed FROM the emitted bytes at compose time and never
re-derived from the JSON, so the worst an implementation change costs is
one spurious "changed" revision after a runtime upgrade. Snapshot
determinism is unaffected — compression wraps the final bytes and reorders
nothing.
Tests: gzip/uncompressed parity through both the instance parses and the
statics, corrupt-gzip leniency through the wrappers, the size guard (with a
falsifiability control proving the same bytes inflate fine under the
production limit), and composer-side blob-starts-with-1F8B +
RevisionHash == SHA256(blob) with a control proving the hash is NOT over
the JSON.
Every public entry point on DeploymentArtifact took ReadOnlySpan<byte> and did
JsonDocument.Parse(blob.ToArray()) — a full copy plus a full document build per
call — so one apply paid ~8-11 full parses of the same bytes (the scoped
ParseComposition alone parsed three times: scope, composition, BuildClusterSets).
DeploymentArtifact.Open(byte[]) now parses once into ParsedDeploymentArtifact,
which owns the JsonDocument and exposes an instance method mirroring each static
entry point. The moved bodies operate on Root; BuildClusterSets takes JsonElement
instead of re-parsing the blob; ResolveClusterScope is memoized per node id, since
every consumer of one apply asks about the same node.
Every static is now a thin wrapper (Open + delegate + dispose) keeping its
ReadOnlySpan<byte> signature and its lenient empty/malformed contract — the
wrapper pays the one copy and owns the JsonException catch, so output shapes are
unchanged. Hot callers migrate to Open() separately.
ParsedDeploymentArtifactTests pins same-bytes-same-output through both doors for
all six entry points over a rich multi-cluster artifact (raw chain + UNS refs +
scripts + keys + ACLs), with falsifiability controls proving the fixture actually
scopes, plus memoization and after-dispose-use coverage.
Accumulated fast-follows from the Task 1-11 reviews. No behaviour change on the
value path.
- Strengthen OpcUaPublishActorBatchTests.Sink_throw_is_swallowed_and_logged. The
survival assertion alone did NOT prove the catch exists: Akka's default
supervision RESTARTS a throwing actor and the restart preserves the mailbox, so
the follow-up message landed either way and the test passed against a build with
no catch at all. It now asserts via EventFilter that the exception NEVER reaches
supervision (Expect(0) on the InvalidOperationException) and that the handler's
own Warning IS logged, keeping the survival assertion. Everything that must have
happened goes inside the Expect(0) action — it counts matches at the end of the
action, so an assertion left outside would race the actor and pass vacuously.
Verified falsifiable: deleting the catch turns the test red.
- MapKey(NodeId): the non-string identifier branch collapsed to a string.Empty key
silently. Such a key cannot match anything the manager stored, so the caller sees
an unexplained miss rather than an error. It now logs a Utils.LogWarning naming
the seam. Every current call site is provably string-only, so this should never
fire.
- The two WireAlarmNotifiers Utils.LogInfo sites interpolated the RealmKey struct,
rendering a raw struct dump. Interpolate .Id and .Realm as separate placeholders
so log-grep patterns stay consistent with the file's other sites.
- Stale comments: the _eventNotifierSources key is a RealmKey now, not a
namespace-qualified NodeId string; and three comments still named
DriverHostActor.ForwardToMux, which Task 8 replaced with ForwardBatchToMux.
Replaces the temporary per-item unwrap bridge with a real batched handler: one
AttributeValueBatchPublished burst now costs ONE hop to the dependency mux and ONE
OpcUaPublishActor.AttributeValueBatchUpdate, instead of one message per value.
Routing per item is unchanged — the v3 Batch 4 single-source fan-out (raw NodeId +
every referencing UNS NodeId, identical value/status/timestamp), with an unmapped ref
Debug-logged and dropped without disturbing the rest of the burst. A burst in which
nothing resolves sends no publish message at all rather than an empty batch.
Indexed for-loop over Items is deliberate: no interface-enumerator allocation per
burst on the hottest routing seam. ForwardToMux and its registrations stay for the
comms-loss fan and per-tag drivers.
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.
Found on the docker-dev live gate, not in tests. Two drivers that faulted
during initialise (rtu-gate, opcua1) each rendered "0 tags" on /hosts against
two authored tags apiece, because a ledger that had never committed returned
SkippedTagReport.Empty.
A zero report is a positive claim -- "this driver skipped nothing" -- and
making it on behalf of a driver that never got as far as looking is exactly
the quiet inaccuracy #541 exists to remove. GetSkippedTags() and
SkippedTagLedger.Snapshot() now return null until the first committed build,
so the column renders the same em-dash it already used for a driver that
cannot report at all. Both are 'no answer'; neither is 'nothing skipped'.
Knock-on: a FOCAS driver whose only init threw now reports null rather than a
zero report, which is strictly more honest than the assertion the test
originally made.
A driver builds its RawPath -> definition table at init and tolerates a
per-tag miss: an unmappable TagConfig, an unparseable address, a
capability-matrix rejection. That tolerance is right -- one bad tag must not
take down the other few thousand. What was missing is any statement of the
fact afterwards. A skip is not an error and happens before any read, so the
driver stays Healthy, errors/5 min stays 0, Last error stays empty, and Last
read advances live off the tags that did map. On the wonder Makino in #540
that reading held while focas-z-34184 ignored 20 of its 22 tags.
New optional capability ISkippedTagReporter, backed by a shared
SkippedTagLedger so the reset/commit/cap rules are one implementation rather
than one per driver. The report -- authored count, exact skipped count, and a
capped sample of RawPath + reason -- rides DriverHealthChanged to a new Tags
column on /hosts, over the Phase 5 telemetry stream. No table: a driver-only
node has no ConfigDb to write to, the lesson from #521.
The issue named eight drivers. It is twelve -- every driver in
DriverTypeNames.All has an authored-tag skip path. Calculation (no scriptId),
MTConnect (no readable DataItem id), Galaxy (no attributeRef) and OpcUaClient
(no nodeId) were unlisted, and OpcUaClient's was the quietest of the twelve:
NamespaceMap drops such an entry with a bare `continue` and not even a log
line. SkippedTagReporterFleetTests pins the invariant reflectively over every
IDriver in src/Drivers/, so a thirteenth cannot re-introduce the silence.
Sql drops a tag at TWO points and its report is the union of both: a blob
that does not map, AND an identifier the 8.1 catalog gate rejects against the
live catalog. The second is what schema drift produces -- a renamed column, a
dropped view -- months after a deployment sealed green, and reporting only
the first would have that driver claim "0 skipped" while polling half its
tags, which is worse than reporting nothing because the column would then be
actively reassuring.
Deliberately additive to driver state, never a change of it. Latching
non-Healthy for a config problem would satisfy both halves of the #527
comms-loss conjunction -- 90 s without a successful read AND non-Healthy --
and a driver whose tags were ALL skipped has no successful reads by
construction, so it would reconnect-loop forever against a healthy device.
Three details that are load-bearing rather than incidental:
- The skip digest is part of the publish fingerprint. A skipped tag leaves
every other fingerprint component identical to a clean driver's, so without
it PublishHealthSnapshot's dedup swallows exactly the publish carrying the
signal -- the trap that already bit the rediscovery and host-status
channels. The digest covers the sample, not just the counts: a redeploy
that fixes one tag and breaks another leaves both counts unchanged.
- The count is exact; only the sample is capped, and the tooltip says so when
it is showing fewer than it counted. A truncated list that looked complete
would have an operator fix twenty tags and believe the job done.
- The ledger stages and commits, so an init that throws part-way leaves the
previous report standing. A partial snapshot would read as "we now skip
fewer tags", the opposite of the truth.
The proto field needs no companion presence flag, unlike host_statuses: a
singular message field carries real presence in proto3, and a has_ flag would
collide with protoc's generated HasSkippedTags.
Verified red-first three ways: dropping SkippedTagDigest from the fingerprint
fails the two dedup tests, removing the seam from one driver fails the fleet
guard, and dropping the catalog-gate leg fails the Sql union test. Full
solution builds with zero new warnings against 4e22f024.
Both defects were found by #491's live gate against a real HistorianGateway and filed rather than
fixed at the time, because each one's fix is a design decision rather than a patch.
#537 — both nodes of a redundant pair captured every value, so the historian received duplicate
live samples for every historized tag (measured: the gateway's write counter advanced twice per
value and halved when one central was stopped).
The recorder now reads a capture gate per value, fed from the new
PrimaryGatePolicy.ShouldRecordHistorizedValues through IHistorizationRoleView — a third verdict on
the same RedundancyRoleView object, registered unconditionally on driver nodes rather than
alongside the alarm section's IRedundancyRoleView (which only exists when that section is enabled;
piggybacking is the #511 registration accident).
Two things are deliberate and easy to get backwards:
* It gates the CAPTURE, not the drain — the opposite of the alarm sink. That queue is in LocalDb
and replicates, so the Secondary already holds the Primary's rows. The historization outbox is
a node-local FasterLog replicating to nobody, so a drain-only gate would leave the standby
appending forever to a queue no one empties: unbounded disk, then a flood of stale duplicates
the moment it is promoted. Entries a node already captured are still drained by that node.
* The peer it stands down for must RECORD THE SAME REFS, not merely hold the same rows — a
Primary in this node's own per-cluster mesh. A node carrying no cluster-scoped role cannot
establish that and never stands down: under a fleet-wide election the elected Primary is
usually running a different cluster's configuration, and deferring to it would end this
cluster's historization silently, with the drain reporting success on an empty outbox forever.
That is the scope mismatch that once suspended the alarm drain on every node of the rig.
An unknown role captures, matching the alarm drain — a duplicate sample is recoverable, a hole in
the trend is not.
#538 — the drain writes every per-tag group and then acks all-or-nothing (deliberate: RemoveAsync
truncates THROUGH an id, so partial acking could drop older unwritten entries). One permanently
failing tag therefore pinned the FIFO head while every healthy tag in the batch was re-submitted
to the historian on every pass, indefinitely — duplicate samples and an outbox that never
truncated, observed live at 8.3 MB from a single unprovisionable tag.
The drain now tracks consecutive failures per historian tag and abandons that tag's entries once
it exceeds MaxWriteAttempts (new, default 5), so the head advances. The counter only advances on a
pass where some OTHER tag wrote successfully: that is what separates "this tag is poison" from
"the backend is down". A count-only quarantine looks identical on the poison-tag case and silently
discards the whole backlog during an ordinary gateway restart, turning a duplicate-samples bug
into a lost-history one. Abandonment is also deferred until every failing tag is over budget, so
nothing is discarded on a pass that could not have advanced the head anyway. Each abandonment logs
a Warning naming the tag; the total rides RecorderStatus.AbandonedEntries.
Capacity also gains a non-zero default (500,000 entries) as the disk backstop for the one case the
quarantine deliberately cannot resolve — a historian refusing every write, where retaining is
correct. Set 0 to restore unbounded retention.
Every new test was falsified by reverting the production change it guards, and each one reddens
alone:
* drop the backend-reachable evidence -> only A_backend_that_fails_every_tag_abandons_nothing
* never abandon -> only A_permanently_failing_tag_is_abandoned_so_healthy_tags_stop_being_rewritten
* latch the capture gate at construction -> only The_capture_gate_is_re_read_per_value...
* treat any Primary as a scope peer -> only the two scope-mismatch cases
Full suite green. (Four OpcUaClient.Browser failures during the run were the opc-plc fixture being
down on the shared Docker host; brought it back up and they pass.)
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
Makes the /raw device tree addressable by an ACL rule, and builds the
scope-path producer PermissionTrieBuilder has accepted since it was written
but which had no caller anywhere.
The plan for this stage said "add NodeHierarchyKind.Raw, whose path is the
RawPath segments Folder/Driver/Device/TagGroup/Tag". That shape is wrong.
RawFolder.ParentRawFolderId and TagGroup.ParentTagGroupId are both
self-referencing, so folders and tag groups each nest without bound and a real
RawPath is folder…/Driver/Device/group…/Tag. NodeScope was fixed-arity and
PermissionTrie.CollectMatches was five hardcoded steps, so the scope model
could not address a real raw tree at all. NodeScope is now {ClusterId, Kind,
Segments} and the walk is a segment loop. The trie STRUCTURE was already
variable-depth (NodeAclPath.Segments is a list, Children is a recursive
dictionary), so only the walker and the scope record had to change.
Four decisions worth keeping:
- The namespace level became the realm. Level 1 gated on NodeScope.NamespaceId,
but v3 has no Namespace entity — the snapshot was retired when the address
space went dual-namespace, and the identifier appears nowhere in production
outside NodeScope. So the level could only ever fail to match, which silently
made every sub-cluster grant unreachable while cluster grants kept working.
It is now a realm token derived from the scope's Kind, not carried on it, so
the walk and the producer cannot disagree about its spelling.
- Provenance comes from the granting row's own ScopeKind, recorded on TrieGrant
at build time. Depth-derived provenance could not name a raw level at all,
since the depth of "Device" varies with folder nesting.
- The trie comparer moved from OrdinalIgnoreCase to Ordinal. Raw sibling-name
uniqueness is ordinal, so Speed and speed can both exist under one device and
a grant on either would have matched both.
- The null-scopePaths "deterministic test mode" is deleted. It planted
unresolvable rows at a direct child of the trie root, a position the walker
cannot reach, so the grant looked placed and could never match — and every
test using that mode proved nothing about production placement. scopePaths is
now required, and a row that does not resolve is dropped with a diagnostic.
A scope resolves to a LIST of paths: a raw tag projected into N equipments is
reachable at N+1 NodeIds and a write through any of them lands on the same
driver reference, so a Tag-scoped grant covering only the raw node would be
bypassable through a UNS one. The lookup is keyed on (ScopeKind, ScopeId), not
ScopeId alone — ScopeId has no FK and no type discriminator, so an id collision
across entity types would hand a grant the wrong node's paths.
Also closed here, all found by review rather than asked for:
- NodeAcl rows had no deploy-time validation of any kind and were not reachable
from DraftValidator at all (DraftSnapshot carried no NodeAcls). They are now
gated for dangling ids, cross-cluster scopes, blank/contradictory ScopeId, the
retired FolderSegment kind, and undefined ordinals.
- ConfigDbApiKeyStore conferred roles from Cluster-scoped grants authored in ANY
cluster, because it filtered on scope kind with no ClusterId predicate while
ArtifactApiKeyStore read an already-cluster-filtered snapshot. On a fused
admin+driver node that was a cross-cluster over-grant. Its doc comment claimed
the two stores matched.
- The two AdminUI ACL banners still said enforcement was LDAP-group -> role
mapping, which stages 1/2 made false in both directions.
- The /apikeys page was never classified in PageAuthorizationGuardTests, so that
test has been red since the key-minting commit merged — I did not run
AdminUI.Tests before merging stage 4.
Falsified rather than assumed: 8 mutations against the trie tests, 5 against the
deploy gate (including one proving a cross-cluster row is not silently
mis-diagnosed as unresolved), a staged red->green for the over-grant regression,
and the realm gate, the comparer and the malformed-path guard each broken
individually to confirm only their own tests go red.
Node ACLs are still not ENFORCED — nothing constructs a NodeScope or calls the
evaluator yet. That is stages 5 and 6; this stage makes the model able to
express what those will enforce.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
A driver-only node has had no ConfigDb since per-cluster mesh Phase 4, so it
registered NullOpcUaUserAuthenticator and denied every UserName token. Its keys
now ride the deployment artifact instead.
- ConfigComposer snapshots ApiKeys alongside the NodeAcls it already carried.
Only the peppered HMAC ships, never a secret, and the pepper is node config
that is deliberately not in the artifact.
- DeploymentArtifact.ParseDataPlaneAuthority decodes both node-side. Grants are
scoped to the node's own cluster; keys are not, because ApiKey carries no
ClusterId — a key is a fleet-wide credential and the ACL rows are what bind it.
- OpcUaPublishActor installs them on DataPlaneAuthorityState on every apply,
BEFORE its plan.IsEmpty early-return: an artifact that only revokes a key
produces an empty address-space plan, so feeding after that return would leave
the revoked key working until an unrelated structural change redeployed.
- ArtifactApiKeyStore serves them to session activation; a node with a ConfigDb
keeps ConfigDbApiKeyStore, whose per-request read is what makes revocation
immediate there.
- OtOpcUaNodeManager gains DataPlaneAuthority — the cluster identity blocker #4
named. NodeScope.ClusterId is required and the evaluator fails closed on a
mismatch, so until now no scope it built could have evaluated. Nothing reads it
until stages 5/6.
The generation is a node-local COUNTER driven by the RevisionHash, not derived
from its bits. PermissionTrieCache treats the highest GenerationId as current, so
a generation taken from hash bytes would install a lower number about half the
time and the cache would go on serving the previous deployment's ACLs, silently.
Sessions are not stamped with it yet — that belongs with the first seam that
reads a UserAuthorizationState, in stage 5, rather than being built now and
consulted by nothing.
Live-gated on docker-dev against site-a-1 (driver-only, no ConfigDb): key
authenticates, wrong secret denied, both halves of the revocation asymmetry
observed, and a SITE-A grant confirmed not to reach site-b-1. Key minted with an
independent Python HMAC so the verifier was checked against something other than
itself. Rig left clean.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
DriverInstanceActor.DisconnectObserved had no production sender. The only paths
into Reconnecting were a failed *initial* connect and an operator's Force
Reconnect, so a driver that connected and then lost its device stayed Connected
indefinitely: publishing Bad for every tag, never re-initialising, and NOT
recovering when the device returned. PublishHealthSnapshot read GetHealth() for
reporting only and never acted on a driver reporting itself unhealthy.
Both comms-loss quality paths -- #477's native-condition annotation and #481's
scripted-alarm input fan-out -- hang off ConnectivityChanged, so both were
correctly wired to a signal that never fired.
The Connected health tick now evaluates a watchdog and self-sends
DisconnectObserved when the driver reports a non-Healthy state AND its
LastSuccessfulRead has been frozen for DefaultCommsLossTimeout (90s).
The conjunction is the design, and the issue's own proposal -- "non-Healthy for
N consecutive polls" -- is wrong. Degraded means *connected but some data is
bad*; several drivers write health per tag inside the read loop
(ModbusDriver.ReadAsync), so one unreadable address latches Degraded forever
against a perfectly reachable device, and a state-only watchdog would reconnect
that driver every 90s indefinitely -- worse than the bug. LastSuccessfulRead is
the honest evidence: preserved across Degraded/Faulted writes, advanced by any
success. The state check remains the second conjunct so an idle Healthy driver
with nothing subscribed is not declared dead for not reading.
Verified across all 12 drivers that a sustained failure does reach a non-Healthy
state -- nine poll drivers via Degraded, MQTT via live connection state, Galaxy
via its reconnect supervisor. CalculationDriver reports Healthy while running
and so can never trip, which is correct: it has no transport.
Also: the Reconnecting retry now calls ShutdownAsync before InitializeAsync.
ModbusDriver (_transport) and MqttDriver (_connection) overwrite an
IAsyncDisposable field inside InitializeAsync without disposing the previous
one, so re-initialising without a teardown leaks a socket per attempt -- latent
while only a manual Force Reconnect reached this path, but the watchdog makes it
automatic and unbounded (~360 leaks for an hour-long outage at the 10s retry).
Tests: 8 new in DriverInstanceActorCommsLossWatchdogTests, each absence
assertion carrying a positive control (a proven count of processed health-poll
ticks). Falsified five ways -- dropping either conjunct reds exactly the test
written for it, and removing the teardown, the null baseline, or the watchdog
call each reds its own guard.
Live gate PASSED on docker-dev (Modbus + otopcua-pymodbus-standard), run twice:
- fixture stopped -> watchdog fires at ~108s, "declaring comms lost"
- fan-out observed at Debug: "pushed BadNoCommunication for 2 mux ref(s)"
(#481 Layer 4) and the #477 native-condition annotation on a driver
owning conditions
- fixture restarted -> reconnected in ~8s with NO node restart; both the raw
(ns=2) and UNS (ns=3) NodeIds read Good with fresh timestamps
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
Layer 4 of the condition-Quality chain (#477 -> #478). #478 gave a scripted
condition the worst quality of its input tags, but only for inputs whose driver
PUBLISHES a Bad data change. A poll driver whose device goes unreachable
publishes nothing — it raises ConnectivityChanged(false) and goes silent — so
the engine held its last Good value, WorstInputStatus stayed Good, and worse,
AreInputsReady stayed true, so the predicate went on evaluating a stale value
after the source had died.
`DriverHostActor.FanCommsLossToDependencyMux` pushes a synthetic
BadNoCommunication with a null value into the dependency mux for every value ref
the disconnected driver owns.
The issue asked for a driver -> all-tag-refs map to be built for this. It
already exists: `_nodeIdByDriverRef`, keyed (DriverInstanceId, RawPath) and
rebuilt from the composition on every apply. Alarm-bearing tags live in the
separate `_alarmNodeIdByDriverRef` and are deliberately excluded — a native
condition is not a mux value and takes the Part 9 quality path instead.
Nothing is pushed on reconnect, deliberately: publishing Good would assert a
value quality this actor has not observed. Recovery rides the driver's own
republish, and the mechanism was verified rather than assumed — the
Connected -> Reconnecting edge calls DetachSubscription, which nulls the
subscription handle, so the reconnect's ResubscribeDesired builds a FRESH
subscription whose PollGroupEngine initial push runs with forceRaise: true.
Without that detach the poll engine would compare against its retained
LastValues, see an unchanged value, raise nothing, and strand these refs Bad
forever — the value-dedup trap that bit the v3 Batch 4 virtual-tag reassert.
Closes#478 Finding 3, which the above depends on. WorstInputStatus skipped
every input whose VALUE was null as "no data yet" — and Bad-with-no-value is
exactly the shape a comms-lost driver produces, so the skip was hiding the case
that matters most; without this half the fan-out above would be silently inert.
The upstream source now stamps a never-pushed path BadWaitingForInitialData —
the honest Part 4 code for that state — and the engine skips that specific code
rather than every null value.
Deleting the null-skip outright is the obvious fix and is wrong in the other
direction: it makes every not-yet-published input drag its condition Bad at
deploy, flooding the quality path with load-time annotations, which is what the
skip existed to prevent. Both mistakes are pinned by a test pair that goes red
under the opposite error, and both were falsified by reverting the production
code:
old null-skip restored -> A_pushed_Bad_carrying_no_value RED
null-skip deleted outright -> An_input_that_was_never_pushed RED
fan-out call removed -> the three DriverHostActor tests RED (1 of 7)
The never-pushed test was vacuous on first writing and is worth recording: the
quality bucket only emits on CHANGE, so the load-time evaluation is the one that
flashes Bad, and subscribing to OnEvent after LoadAsync made the assertion
unfalsifiable — it passed against the broken behaviour. It now subscribes first.
`FakeUpstream.ReadTag` returned 0x80340000 (BadNotConnected) for an unset path,
a different Bad code from production's placeholder. Harmless while any Bad
behaved alike; load-bearing now, so it is aligned and pinned by a reflection
guard asserting the engine's private constant agrees with what the production
source stamps.
Behaviour change to weigh, flagged in the issue and confirmed: virtual tags
computed from a comms-lost input now go Bad too. That follows from using the
shared mux and reads as correct, but it is broader than scripted alarms.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
Both `IDriver` members had no production call site. Their only consumer was
`MemoryTracking`, deleted with the driver-tier recycle machinery in #522, and
`GetMemoryFootprint`'s doc-comment claimed "polled every 30s by Core" — which
had not been true since v3, even before #522 removed the poller.
Owner decision on #525 was (a) delete rather than (b) build the process-wide
watchdog that would consume them. The driver-side implementations are in git
history and were mostly reusable; what was always missing is the trigger, and a
member nobody calls is how the misleading doc-comment survived this long.
Removed from `IDriver`, all 12 driver implementations, and every test stub —
plus the support code that existed only to feed them, which is the part a
member-only deletion would have left behind as dead weight:
- MTConnect's three per-entry byte-weight constants, its cached
`ProbeModelDataItemCount`, and `CountDataItems` — a full device-tree walk run
on every cache fill to populate a number with no remaining reader.
- OpcUaClient's `_discoveredNodeCount` field and the discovery-pass write to it.
- MQTT's `ApproxBytesPerAuthoredTag`.
Two behavioural notes, neither a change in what the software does:
- `AbCipTemplateCache` now has no production reset path. Its only caller was
AB CIP's flush, which itself had no caller — so nothing changes except how
visible the fact is. Recorded on the type, along with the correction that its
comment's claim about a device `ReinitializeAsync` also flushing it was never
true. A UDT redefined on a live controller keeps its stale shape until the
driver instance is rebuilt.
- MTConnect's `/probe` cache is still dropped, by `InvalidateProbeModel()` on an
Agent restart. That is now `internal` rather than `private` so the tests which
used the flush purely to empty the cache drive the real path instead — they
now cover a case that actually occurs in production. Several unrelated doc
comments cited `FlushOptionalCachesAsync` as the canonical CAS example and are
repointed to it.
Tests: the ones that existed only to assert on the removed members are deleted
(3 OpcUaClient, 3 Galaxy, 2 MQTT, 1 MTConnect, 1 AbCip). Worth noting one was
vacuous — `FlushOptionalCachesAsync_resets_footprint_counter` asserted a
footprint of 0 before the flush and 0 after with no discovery in between, so it
would have passed against a flush that did nothing at all.
Full suite green except 7 failures verified identical on master before and
after (4 AbLegacy integration with the compose profile down, 3
Host.IntegrationTests), checked by name rather than assumed.
Docs corrected where they described this as live: `DriverLifecycle.md` (which
carried the #525 placeholder), `drivers/AbCip.md`, `drivers/AbLegacy.md`, and
`v2/driver-stability.md` — whose "polls this every 30 s" + per-driver
`DriverConfig` memory budget were never built as written and are now banner-ed
as design intent rather than description.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
#519 — a driver's OPC UA status code now reaches a client with its sub-code intact.
DriverInstanceActor projected the driver's uint onto a 3-state Good/Uncertain/Bad enum
and OtOpcUaNodeManager re-expanded Bad to a bare 0x80000000, so BadNoCommunication,
BadTypeMismatch and BadNotSupported were indistinguishable on the wire and every
driver's status constants (incl. the 16 #497 had just corrected) were diagnostic in
logs only. Resolved by option (b): the raw uint rides AttributeValuePublished ->
AttributeValueUpdate -> IOpcUaAddressSpaceSink.WriteValue, OpcUaQuality is deleted, and
severity classification is a helper (OpcUaStatus.IsGood/IsUncertain/IsBad).
The surface was wider than the issue costed (41 files, not "the sink and 3
implementations"): OpcUaQuality had non-driver producers with no uint to carry.
Each got its natural code — BadInternalError for a failed virtual-tag script,
BadNoCommunication for driver comms loss — and ScriptedAlarmHostActor's pair of
mutually-inverse uint<->enum mappers is deleted, having round-tripped a code the
engine already held. AddressSpaceApplier's removal writes deliberately keep the
generic Bad: not losing a code the driver had is this issue; inventing one for the
deploy path is a separate call.
Behaviour change: WriteAlarmQuality fires its Part 9 event on a status-code change
rather than a severity-bucket change. While the annotation was 3-state those were the
same predicate; now a move between two Bad reasons changes what a subscriber reads.
New SDK-free named constants live on Commons.OpcUa.OpcUaStatus, and
StatusCodeParityTests is widened to cover them — that guard is reflection-driven, so
it can only see a named constant, never an inline literal.
#517 — the symptom is real but the issue named the wrong path and its fix was unsafe.
ShutdownAsync cannot race: DriverInstanceActor calls it from PostStop, after the
mailbox drains. (Core/Hosting/DriverHost, the one class with a concurrent GetHealth,
has no production call site.) InitializeAsync can: it runs on a fire-and-forget
Task.Run, and MTConnectDriver writes Healthy then starts loops that can throw into a
catch which disposes the client before writing Faulted.
Moving the write earlier, as suggested, inverts the terminal state — teardown drains
in-flight loops that write health as they unwind, so a late loop could leave Healthy at
rest with every connection closed. The fix is a bracket (conservative before, terminal
after), applied to MTConnect's init-failure path only; shutdown keeps its single
trailing write, and the rule plus the reachability table are documented on
IDriver.ShutdownAsync.
CalculationDriver.GetHealth returned a hard-coded Healthy that consulted nothing —
Healthy before initialize, after shutdown and after dispose. Now derived from _running.
Its test, named Health_is_always_Connected, asserted the defect and is replaced.
Every new guard was falsified by breaking the production code and watching it go red.
The two sub-code tests each assert two different Bad codes, since a single-code
assertion could pass against a hard-coded constant under the old projection.
Pre-existing failures, verified unchanged on a stashed tree and not chased: 4 in
Driver.AbLegacy.IntegrationTests (slc500/micrologix/plc5 fixtures not up) and 3 in
Host.IntegrationTests.
Closes#517Closes#519
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
`IHostConnectivityProbe` was a dead surface: eleven drivers implement it, `GetHostStatuses()`
had ZERO production call sites, and `OnHostStatusChanged` had no subscriber outside the Galaxy
driver's own aggregator. Per-host connectivity was computed by every driver and read by nobody.
The issue offered "build the publisher or delete it". The publisher as its entity doc described
it — driver nodes upserting `DriverHostStatus` rows — is not buildable: per-cluster mesh Phase 4
gates `AddOtOpcUaConfigDb` on the `admin` role, so a driver-only node has no ConfigDb connection
to write rows with. So the capability is kept and the transport changed.
`DriverHealthChanged.HostStatuses` now carries the probe result to `/hosts` as a Hosts column.
That channel already reached the page, already survives the mesh split via the Phase 5 gRPC
telemetry stream, and already replays a last-value snapshot on re-subscribe — so per-host state
re-primes after a reconnect without a durable store. Both halves of the interface finally do
what they are for: the event triggers a prompt publish, the pull is the source of truth.
The point of the column is the case the driver-level state chip structurally cannot express: a
multi-device driver stays aggregate-Healthy while ONE of its devices is unreachable.
Two traps, both pinned by tests that were falsified against the prod code:
- The host digest MUST be in the publish fingerprint. On a single-host-down transition every
other fingerprint component is unchanged, so the dedup would swallow exactly the publish
carrying the news — the trap that already bit the rediscovery signal. Removing it turns the
guard test red, verified.
- null (no probe) must stay distinct from empty (probe with no hosts). proto3 cannot tell an
absent repeated field from an empty one, hence the explicit `has_host_statuses` flag;
collapsing them would render every probe-less driver as one whose devices are all fine.
Dropped: the DriverHostStatus entity, enum, DbSet, model config and table (migration
DropDriverHostStatusTable — empty on every deployment, so the scaffolder's data-loss warning is
moot, and Down() recreates it exactly).
Found en route, NOT fixed here: `DriverInstanceResilienceStatus` is the identical defect — no
writer, no reader, only a DbSet declaration, while the live data rides the
`driver-resilience-status` telemetry channel. Its doc-comment now states that rather than
describing the sampler and AdminUI join that were never built. Filed as #524 rather than widening
this schema change beyond what was asked.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
tag gate, and fix three silently-broken OpcUaClient integration tests
Continues working through deferment.md's remaining items.
SKIPPED TESTS (13 -> 0 for this reason). The skip said "dark until Batch 4";
Batch 4 shipped as v3.0 and RETIRED the equipment-tag path rather than lighting
it, so every one of them was waiting on something that had already happened and
gone the other way. Resolved individually rather than in bulk:
- REVIVED onto the raw/UNS shape (6): four DriverHostActor primary-gate cases,
the cluster-scoping rebuild, and the real-SDK dual-namespace materialisation
E2E. The behaviour never went dark, only the v2 seeding did.
- FOLDED into a live parity test (3): DeploymentArtifactRawUnsParityTests
already compared RawTagPlan record-equal and RawTagPlan carries array /
historize / alarm intent, so widening its corpus by two tags covers what
three empty placeholder suites had been promising. Those suites are deleted.
- DELETED (4): subjects genuinely retired -- equipment-namespace EquipmentTags,
and both dot-joint {{equip}}.X token suites (that resolver was deleted in
Batch 3; the slash-joint form is covered by DeploymentArtifactEquipRefParityTests).
Falsified, not assumed: removing the seeded UnsTagReference turns two revived
primary-gate tests red; re-homing the SITE-A node to MAIN turns the scoping test
red. Widening a parity corpus also needed direct value assertions -- parity alone
cannot distinguish "both seams right" from "both seams blind".
Runtime.Tests skips 13->3, OpcUaServer.Tests 4->1; all remaining are deliberate
EquipmentDeviceBindingRetired tombstones.
G-7 (deploy-time TagConfig gate). MQTT shipped an Inspect() nobody ever called;
now wired. Replaced the hand-maintained list with a reflection guard, because the
existing test enumerates driver types by hand and so guards renames but can never
notice a gap. NOTE: the first version of that guard was hollow -- it discovered
candidates via GetReferencedAssemblies(), which is circular, since the compiler
elides a reference nothing in the IL uses. Removing MQTT from the map also removed
its Contracts assembly from the manifest, and the guard passed green with the bug
reintroduced. Rewritten to scan the output directory; re-falsified and it now
fails naming MqttTagDefinitionFactory. Residual recorded at the code: Sql,
MTConnect, Calculation and OpcUaClient have no Contracts-side Inspect at all, so
the deploy gate stays weaker than the AdminUI's authoring gate for them.
OPCUACLIENT INTEGRATION TESTS (3 of 4 failing, on master too). Not fixture rot:
the simulator was healthy and Client.CLI read the very same node Good. v3 made the
driver resolve every reference through its authored RawTags table, so a bare
"ns=3;s=StepUp" fails LOCALLY at the resolver with BadNodeIdInvalid and never
reaches the wire -- the tests were exercising the unresolved-reference guard.
Fixed by seeding OpcPlcProfile.RawTags in the deploy artifact's TagConfig shape
and addressing by RawPath. 4/4 pass.
DOCS. Applied the Phase7* -> AddressSpace* rename across the four live docs
(historical bannered files deliberately untouched), resolving the three that are
not renames to their real members. Found en route that the earlier banner pass
missed three files: VirtualTags.md, OpcUaServer.md and ScriptedAlarms.md all
still presented the test-scaffolding GenericDriverNodeManager as the production
dispatch path. All three now carry the correction.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
Removes the tag-set drift detector shipped in e77c8a35. The design was sound
-- the tautology drivers were correctly gated out, and the structurally
undetectable "vanished" direction was declared rather than faked -- but none of
that is the objection. The feature browsed real plant equipment on a recurring
timer, at a frequency no operator could configure (the interval was a
constructor parameter with no appsettings key), and that outbound traffic was
never once observed against a real PLC, CNC or MTConnect Agent.
Removed: TagSetDrift/TagSetDriftDetector; DriverInstanceActor's drift timer,
tick message, Props/ctor parameter, gate, handler and leaf-collector;
ITagDiscovery.AuthoredDiscoveryRefs + .DiscoveryStreamIncludesAuthoredTags and
their four implementations; 14 tests.
Kept: ITagDiscovery.SupportsOnlineDiscovery (it predates this and drives the
universal browse picker), and the whole IRediscoverable consumption from
09a401b8 -- the driver tells us, we do not poll.
CONSEQUENCE, deliberately accepted: AbCip and FOCAS browse a live backend but
implement no IRediscoverable, so they now have NO tag-change signal at all -- a
PLC re-download is invisible until someone re-browses by hand. Recorded in
CLAUDE.md so the next reader does not have to rediscover it, along with the
shape to reach for if that coverage is wanted back (event-driven, or triggered
by a reconnect/deploy, not a wall-clock timer).
Runtime.Tests 476 pass / 13 skipped (skip set unchanged); FOCAS 275, TwinCAT
192, MTConnect 491, AbCip 342, Core.Abstractions 266 all green.
Claude-Session: https://claude.ai/code/session_015p7wGqy3YpZNCpDzTpGMKo
The second thing #516/§8.2 deliberately did not build. The original objection
stands — Modbus, S7, MQTT, AbLegacy and Sql echo authored config from
DiscoverAsync, so diffing discovered against authored is a tautology for them
— but it is now ENCODED rather than used as a reason not to build.
The discriminator already existed: ITagDiscovery.SupportsOnlineDiscovery is
documented as "enumerates the tag set from the live backend rather than
replaying pre-declared/authored tags", and it separates the fleet cleanly —
FOCAS, TwinCAT, MTConnect and AbCip true; the five config-echo drivers
explicitly false. The check runs only for a driver that is
SupportsOnlineDiscovery AND reports the new AuthoredDiscoveryRefs. That is
nullable rather than empty-by-default on purpose: an empty collection
legitimately means "nothing authored, everything discovered is new", so
conflating the two would report total drift for a driver that never opted in.
Where the value is: AbCip and FOCAS browse a live backend but implement no
IRediscoverable, so before this they had no change signal at all. A periodic
compare is the only way to notice a PLC re-download.
A finding that changed the design: FOCAS, TwinCAT and AbCip all re-emit their
authored tags into the same DiscoverAsync stream as device-derived ones, so
the browse picker can show an operator their existing tags. Discovered is
therefore always a superset of authored, and a VANISHED tag can never appear
missing — one whole direction is structurally undetectable for three of the
four. Shipping a Vanished list that is permanently empty for them would have
been the half-inert seam this whole exercise removes. So the driver declares
DiscoveryStreamIncludesAuthoredTags, the detector does not compute the
undetectable half, and the operator message says "missing-tag detection
unavailable for this driver" rather than letting the absence of a missing-tag
clause read as reassurance. MTConnect is the only driver where both directions
work — its stream is purely probe-model-derived.
Behaviour: a failed browse is not drift (an unreachable device would otherwise
report every authored tag as vanished); a truncated capture is skipped for the
same reason; an unchanged drift is reported once rather than re-stamping its
timestamp; drift that resolves clears the prompt. Interval defaults to 5
minutes — it browses a real device, and a tag set changing is an engineering
event, not a runtime one. It reuses the Stage-2 surfacing, raising the same
/hosts re-browse prompt, and never rebuilds the served address space.
Comparison logic is a pure, separately-tested type rather than actor-inline.
14 new tests. The gate was verified load-bearing by deleting the
SupportsOnlineDiscovery half: the tautology-driver test goes red, and it
asserts discovery was never invoked rather than merely "no prompt raised",
which would have passed for the wrong reason if the timer never fired.
Full suite green except the same three pre-existing fixture-gated integration
suites, identical counts.
Five drivers ignored the config handed to them on reinitialize, serving the
options their constructor captured — and the deployment sealed green anyway.
Fixed on both halves, as chosen.
Seam (load-bearing): DriverSpawnPlanner routes a changed DriverConfig to
ToStop + ToSpawn instead of an in-place delta, making the factory the single
parse authority. This REVERSES the deliberate decision documented at
DriverSpawnPlan.cs:49-50 ("a pure DriverConfig change stays an in-place delta
— no reconnect"). That reasoning was correct about the resilience pipeline and
wrong about the driver. The accepted price is a reconnect per config edit.
Per-driver, and this split was not anticipated:
- Re-parse in place — Modbus, AbLegacy, OpcUaClient. ParseOptions extracted
from each factory, called from InitializeAsync behind a HasConfigBody guard
so "{}" still keeps the constructor options.
- Respawn-only, deliberately NOT re-parsed — Sql and FOCAS. Each builds more
than options from config (Sql's dialect + resolved connection string,
FOCAS's client-factory backend, both injected at construction), so adopting
new options alone would run a NEW tag set against an OLD connection. Half a
re-parse is worse than none. SqlDriver's doc-comment asserted the opposite
premise — "config parsing belongs to the factory, which builds a fresh
instance" — which was false when written and is true only now.
Two green seals removed from ApplyChildDelta: it overwrote the cached Spec
synchronously BEFORE the child dequeued the message, so the host believed the
new config was live and the next reconcile computed no delta, sealing the
drift permanently; and it Tell'd with no Receive<ApplyResult> registered, so a
failed reinit — including Galaxy's deliberate NotSupportedException —
dead-lettered.
Exposed (not created) by the change: a factory throw is a CONFIG error, and
SpawnChild catches it and silently substitutes a stub. Only a brand-new driver
could reach that before; an ordinary config edit can now. Raised Warning ->
Error with an actionable message. It still does not fail the deployment —
doing so would let one malformed driver block a fleet deploy, so that is a
follow-up rather than a drive-by.
Tests: every pre-existing reinit test in these suites passes "{}", the exact
input a guarded re-parser treats as "keep constructor options" — blind to this
defect by construction. New tests pass CHANGED json; the Modbus one was
verified falsifiable by deleting the re-parse (goes red). Two tests asserted
the old behaviour and were rewritten, including the positive control in
DriverHostActorUnreadableArtifactTests, whose observable moved from
UnsubscribeAsync to ShutdownAsync now that teardown happens by stopping the
child rather than emptying its desired set.
Remaining full-suite failures are pre-existing and environmental, verified
identical on the pre-change tree: Host.IntegrationTests (3) and
Driver.AbLegacy.IntegrationTests (4, docker fixture-gated).
#507 was filed as "injection inert in v3, re-migrate onto the raw subtree".
Reading it, the retained code is not revivable: it resolves equipment from
EquipmentNode.DriverInstanceId UNION EquipmentTags, and BOTH are structurally
empty in v3 (AddressSpaceComposer always constructs EquipmentNode with a null
DriverInstanceId; DeploymentArtifact hard-codes an empty EquipmentTags set).
Removing the guard would have changed a log line and injected nothing. So it
is deleted rather than fixed, and #507 closes as superseded by /raw
browse-commit.
Removed: HandleDiscoveredNodes, PartitionDiscoveredByDeviceHost,
ShouldWarnPartition, PlansRoutingEqual, ApplyDiscoveredPlansForDriver,
_discoveredByDriver, the redeploy re-inject tail, DiscoveredNodeMapper,
DiscoveredInjection, AddressSpaceApplier.MaterialiseDiscoveredNodes,
OpcUaPublishActor.MaterialiseDiscoveredNodes, and the Runtime-local copies of
CapturingAddressSpaceBuilder/DiscoveredNode.
The connect-time discovery loop goes too (StartDiscovery, RediscoverTick,
HandleRediscoverAsync, DiscoveredNodesReady, TriggerRediscovery). With
injection gone it had no consumer, and leaving it would either dead-letter or
keep browsing real devices up to ~15x per connect to drop the result.
ITagDiscovery itself stays — the /raw browse picker drives it through
Commons/Browsing/DiscoveryDriverBrowser, which never went through the actor,
so the picker is unaffected.
deferment.md §4.4 called the 18 DiscoveryInjectionDormantV3 tests "Real —
blocked on #507". They are not: they assert an equipment-rooted graft
(EquipmentRootNodeId == "EQ-1") that v3 cannot produce, so they could never
have unskipped as written. Deleted along with DiscoveredNodeMapperTests, the
CapturingAddressSpaceBuilder tests, and the MaterialiseDiscoveredNodes tests
in AddressSpaceApplierTests/OpcUaPublishActorTests. Runtime.Tests skipped:
31 -> 13.
IHostConnectivityProbe is deliberately NOT resolved here. GetHostStatuses()
has no production call site and nothing writes a DriverHostStatus row, but the
table was re-created in the v3 initial migration, so deleting on inference
would be wrong. Split to Gitea #521 with both options costed.
CLAUDE.md, docs/drivers/{Galaxy,TwinCAT,MTConnect}.md updated — they described
the seam as dead.
Build clean; Runtime.Tests 476 passed / 13 skipped; OpcUaServer.Tests 362
passed / 4 skipped.
Nine drivers raise IRediscoverable.OnRediscoveryNeeded when they observe that
a remote's tag set may have changed — a Galaxy redeploy, a TwinCAT
symbol-version bump, an MTConnect agent restart, a Sparkplug rebirth. Nothing
in src/ subscribed, so every raise went into the void. Drivers even wrap the
invoke in try/catch for "subscriber threw"; there has never been a subscriber.
DriverInstanceActor now attaches the event in PreStart and detaches in
PostStop, mirroring AttachAlarmSource/DetachAlarmSource. Attach is per-actor,
not per-connect: the raise has no connection affinity and can land while the
driver is between connects. The detach is load-bearing rather than tidy — the
IDriver instance outlives the actor when the host respawns a child around the
same driver object, so a missing -= accumulates one handler per respawn, each
holding a dead Self.
The signal rides the existing driver-health snapshot to /hosts as a
"re-browse" chip. It is ADVISORY: the served address space is deliberately not
rebuilt, because v3 authors raw tags explicitly through the /raw
browse-commit flow and a runtime graft would materialise nodes no operator
approved and no deployment artifact records.
Two things worth knowing:
- PublishHealthSnapshot dedups on a health fingerprint, and a rediscovery
raise on an otherwise-healthy driver changes none of the four fields it
hashed. Without adding the timestamp to that tuple the dedup swallows the
exact publish carrying the signal. Reverting that one line turns 3 of the 5
new tests red, which is how it was confirmed rather than assumed.
- The new test stub implements IDriver from scratch instead of deriving from
the shared StubDriver, whose GetHealth() returns DateTime.UtcNow — its
fingerprint differs on every call, so the dedup never engages and the dedup
test would have passed vacuously either way.
The planned discovery-result drift detector was NOT built. Modbus, S7, MQTT,
AbLegacy and Sql all echo authored config from DiscoverAsync rather than
browsing the device, so a discovered-vs-authored diff is permanently empty for
them — it would have been another plausible-looking inert seam, which is the
class this audit exists to remove. IRediscoverable is the trustworthy signal
because the driver asserts it deliberately.
DriverHealthChanged, IDriverHealthPublisher.Publish and HostsDriverRow gain
defaulted optional parameters, so no existing call site changed. telemetry.proto
gains fields 8/9 and both Phase-5 mappers carry them.
Runtime.Tests 512 passed / 31 skipped.
Option A from the issue — the cheap consistency check, no interface change.
The desired ref set was re-applied on a Connected TRANSITION (`ResubscribeDesired`)
and on a deploy's `SetDesiredSubscriptions`, and nowhere else. A subscription lost
while the driver STAYED Connected — a subscribe that failed and was never retried,
a handle dropped down an error path — left the actor subscribed to nothing while
still reporting Healthy, indefinitely.
`ReconcileSubscription` now rides the existing 30 s `health-poll` tick: while
Connected, a driver holding a desired set but no live handle is re-subscribed.
Gated on `ISubscribable`. A driver that cannot subscribe at all can still be handed
a desired set, and without the guard it would be told `Subscribe` every tick and
fail every tick, forever.
This is a state-machine consistency check, NOT a probe — it can only see that this
actor holds no handle, never that a handle is present but dead server-side, because
`ISubscriptionHandle` is opaque. That is option B, and it stays deferred until a
real occurrence shows the handle outliving the subscription; it would need a
liveness member implemented across all eight drivers with different subscription
mechanics.
A redundant `Subscribe` is possible (a tick queued ahead of an already-queued one
observes the same no-handle state) and is harmless: `HandleSubscribeAsync` is a
`ReceiveAsync`, so no tick is processed mid-subscribe, and it drops any prior
subscription before establishing the new one. Suppressing it would need a
pending-subscribe flag threaded through every self-tell site — more state than the
race is worth.
`healthPollInterval` becomes an optional Props parameter, mirroring the existing
`rediscoverInterval` seam, so a test can drive the reconcile without waiting 30 s.
Three tests. The reconcile itself is proven by a stub whose FIRST subscribe throws:
nothing else would ever retry it, so a second attempt can only come from the
reconcile — disabling `ReconcileSubscription` turns it red. The two guard tests are
absence assertions and are bounded by a positive control rather than a sleep: a
counting health publisher makes ticks PROCESSED observable, so "still one subscribe"
means the reconcile ran and declined, not that the timer had not fired.
Explicitly not deploy-time drift (#486) — re-asserting an already-correct desired
set would not have helped there.
`AwaitAssert` returns on its FIRST success, so an assertion that is already true at
t=0 passes instantly and spends none of its duration. All three sites asserted an
absence against a collection that starts empty, so none of them ever gave the
handler the time their comments claimed.
Replaced with mailbox-ordering positive controls rather than sleeps:
- `RouteNativeAlarmAck_unknown_node_is_dropped` — follow the unmapped ack with a
MAPPED one. The host forwards via `Tell` and the child handles `RouteAlarmAck`
with `ReceiveAsync` (which suspends its own mailbox), so once the control reaches
the driver the unmapped one has provably been handled.
- `RouteNativeAlarmAck_on_non_primary_is_dropped` — the role itself is the control:
return the node to Primary and re-send. Comments (`gated` / `control`) are the
discriminator, so a leaked ack is visible rather than merged into a count.
- `PeerProbeSupervisorTests.Single_node_snapshot_spawns_no_children` — a FOURTH site
the issue did not list. It says the other `ChildCount.ShouldBe(0)` waits are each
preceded by a `ShouldBe(1)`; that is true of three of the four, but not this one,
which asserts the initial state. Now followed by a two-node snapshot: the count
reaching 1 proves the single-node snapshot was processed, and a local-node child
would have made it 2.
The issue warned these may turn red, which would be a real defect surfacing. They
did not — the routing is correct. Proven by breaking it deliberately instead:
disabling the Primary gate reddens the Secondary test, and routing unmapped ids to
an arbitrary mapping reddens the unknown-node test. Under the old form both breaks
passed.
The reported symptom was "Runtime.Tests occasionally reports Failed: 1" with no test
name. Instrumenting 30 full-assembly runs reproduced it at 13% (4 runs) and showed it
was never ONE flaky test: five failures across three distinct tests in that batch, and
five more distinct tests over the verification rounds that followed.
Two hypotheses were tested and DISPROVED by measurement before anything was changed:
- Cluster-formation timeout. Every test in this assembly forms a real single-node Akka
cluster over TCP in RuntimeActorTestBase's constructor and waits 5 s for MemberStatus.Up
— 219 formations per run. Instrumented: idle max 960 ms, under-load max 1470 ms, zero
timeouts. A 3.4x margin; not the cause.
- Ephemeral-port exhaustion from that bind churn. TIME_WAIT measured at ~0 throughout.
Not the cause.
FOUR GENUINE ORDERING/LOGIC DEFECTS (none is a timeout):
1. VirtualTagHostActorTests.ApplyVirtualTags_respawns_child_when_plan_changes_in_place
The test loops to accept RegisterInterest/UnregisterInterest in EITHER order — and
then asserts on mux.LastSender, which DEPENDS on that order. Under the [Register,
Unregister] interleaving LastSender is the DYING child and the assertion fails. Now
captures the sender of the Register message as it arrives. Fully deterministic; no
timing involved. (Swept the other four LastSender sites: all drain the Unregister
first, so their ordering is fixed. Only this one was unsafe.)
2. DriverHostActorWriteRoutingTests.Primary_routes_write_via_uns_nodeid_to_driver_by_rawpath
ApplyAck marks the end of the APPLY, not the point a write can succeed: the child is
still in Connecting, which deliberately fast-fails writes ("driver not connected"),
and the NodeId->driver reverse map has not been pushed. Now retries until accepted.
This does not weaken the assertion — every rejection branch replies WITHOUT reaching
the driver, so Writes.Count.ShouldBe(1) still means exactly what it did.
3. HistorianAdapterActorTests.Redundancy_snapshot_without_local_node_...
One 500 ms constant served both presence budgets (AwaitAssert) and absence windows
(ExpectNoMsg). Different quantities that happened to share a number, so the presence
budget could not be raised without slowing every absence check. Split into
AssertTimeout (5 s) and Settle (500 ms).
4. ContinuousHistorizationRecorderTests.Retry_after_writer_failure_eventually_acks
Waited on `outbox == 0`, which is ALSO true before the first append — so the poll
could win the race against the recorder and return having observed the INITIAL state,
after which the CallCount>=2 check outside the block failed against a recorder that
had not run. Both conditions are now polled together with the retry count as the
discriminator. (The preceding test in the same file documents this exact trap.)
THE PRESENCE-BUDGET CLASS:
After those four, three consecutive 30-run rounds each still failed once — on a
DIFFERENT test, in a DIFFERENT file, every time (OpcUaPublishActorApplyFailureTests 2 s,
OpcUaPublishActorRebuildTests 2 s, OpcUaPublishActorTests 500 ms). That is a
distribution, not a defect, and fixing it one test at a time was the wrong method.
RuntimeActorTestBase now carries a documented PresenceBudget (15 s) and 57 sub-3 s
AwaitAssert/AwaitCondition durations across four files route through it.
Why this is safe rather than sloppy: a presence budget is an upper bound before giving
up, not a wait — these helpers poll and return the instant the condition holds. Raising
one costs nothing on the happy path and CANNOT make a genuinely failing assertion pass;
it only changes how quickly a real breakage reports. Absence windows are the opposite
(the elapsed time IS the assertion) and were deliberately left untouched with their own
short, individually-calibrated literals.
ScriptedAlarmHostActorTests is a special case, diagnosed rather than merely raised: its
8 s budget was sized for message passing but actually waits on a REAL Roslyn compile
(ApplyScriptedAlarms -> ScriptedAlarmEngine.LoadAsync compiles every predicate). Cold
script compiles are the slowest and most variable thing in the assembly, so that class
gets 30 s with the reason recorded.
Verification: 30 consecutive full-assembly runs, 30 passed / 0 failed, against a 13%
baseline. Full solution builds; all 41 unit-test projects green.
NOT fixed here, filed as #501: DriverHostActorNativeAlarmAckRoutingTests:90 and :116 use
AwaitAssert for an ABSENCE assertion, so they return instantly and prove nothing. They
are excluded from the budget change on purpose — a bigger number does not fix them, and
rewriting them to settle-then-assert may legitimately turn them red.
Every value verified against Opc.Ua.StatusCodes in the pinned SDK (1.5.378.106) by
reflection, not by copying siblings. All are client-visible: OPC UA clients branch on
status.
The three named in #497:
- Historian.Gateway SampleMapper "BadNoData" 0x800E0000 -> 0x809B0000
(0x800E0000 is BadServerHalted)
- Galaxy "BadTimeout" 0x800B0000 -> 0x800A0000
(0x800B0000 is BadServiceUnsupported)
- TwinCAT BadTypeMismatch 0x80730000 -> 0x80740000
(0x80730000 is BadWriteNotSupported)
BadTypeMismatch was wrong in three MORE drivers the issue did not name — FOCAS,
AbLegacy and AbCip carried the same 0x80730000. Every call site is a
FormatException/InvalidCastException conversion failure, so the name was right and
the value was wrong in all four.
Adding the reflection guard then surfaced nine further defects offline tests had
never checked:
- Galaxy GoodLocalOverride 0x00D80000 -> 0x00960000 (not a UA code at all)
- Galaxy UncertainLastUsableValue 0x40A40000 -> 0x40900000
- Galaxy UncertainSensorNotAccurate 0x408D0000 -> 0x40930000
- Galaxy UncertainEngineeringUnitsExceeded 0x408E0000 -> 0x40940000
- Galaxy UncertainSubNormal 0x408F0000 -> 0x40950000
- TwinCAT BadOutOfService 0x80BE0000 -> 0x808D0000 (was BadProtocolVersionUnsupported)
- TwinCAT BadInvalidState 0x80350000 -> 0x80AF0000 (was BadAttributeIdInvalid)
- AbCip/AbLegacy GoodMoreData 0x00A70000 -> 0x00A60000 (was GoodCommunicationEvent)
- Historian.Gateway GatewayQualityMapper Good_LocalOverride 0x00D80000 -> 0x00960000
Galaxy's whole Uncertain block read as though the OPC DA quality byte could be
shifted into the UA substatus position; it cannot — the substatuses are an unrelated
enumeration. GatewayQualityMapper already held the correct table, which is what the
Galaxy values are now reconciled against.
Guard: StatusCodeParityTests (Core.Abstractions.Tests, which already project-
references every driver) reflects over every `const uint` in the deployed
ZB.MOM.WW.OtOpcUa.Driver.*.dll whose name reads as a status code, and asserts it
equals Opc.Ua.StatusCodes.<name>. Discovery is by convention, so a new driver
assembly referenced by that project is covered with no edit here. It went red on all
nine defects above before the fixes and now checks 109 constants green.
Because reflection can only see a NAMED constant, two inline literals were hoisted so
the guard can reach them: Galaxy's BadTimeout/Bad into StatusCodeMap, and
GatewayQualityMapper's 15-entry table into a new HistorianStatusCodes. An inline
`0x800B0000u, // BadTimeout` at a call site is exactly how the Galaxy defect survived.
The drivers stay SDK-free by design; only the test project references Opc.Ua.Core,
as the oracle.
Also corrected: tests that pinned the wrong values (Galaxy StatusCodeMapTests,
SampleMapperTests, GatewayQualityMapperTests — all written from the same bad source),
a mislabelled comment in DriverInstanceActorTests, and stale constants in two design
docs, one of them the unexecuted MTConnect driver design that would have propagated
BadNoData's wrong value into a new driver.
The CLI's SnapshotFormatter value->name table was checked against the SDK and is
correct; no change needed.
Joseph Doherty