- NodeManagerLockTimeBenchmark: wall-clock budget enforced only with OTOPCUA_BENCH=1; default run keeps the state-correctness asserts and logs timings.
- OpcUaClientDriverWire host-status: signal (semaphore) wait that only accepts a non-Running state observed after the server is disposed.
- AbCip PollLatencyComparison: retry once on a fixture connection reset; p95 criterion enforced only with AB_LAT_ENFORCE=1.
- ScriptRunGuardTests: orphan scripts held by a ctx.Released signal instead of a 250 ms wall-clock spin that could finish inside the 60 ms budget under load; stuck-orphan wall-clock rule disabled in the tests that pin the strike counting.
- DependencyMuxReplayTests: split the inside-grace positive control (1 h grace) from the release test (control asserted before unregister), removing the 200 ms grace race.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
Behaviour-neutral hop: pin the SDK AggregateManager default to our
TreatUncertainAsBad=false (SDK .182 flipped it to true), parity guard now an
explicit deliberate-divergence assertion plus a booted-server check; adapt to
obsolete FindPredefinedNode(Type)/TraceConfiguration; Microsoft.Extensions.
Logging.Abstractions 10.0.7 -> 10.0.8 (SDK floor); stall wire test's late-flush
positive control is best-effort (.176 preserves publish queue ordering).
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
QuestDB modified-history reads dropped every still-pending trail row that was
not the latest at its source time, so V0->V1 (applied marker never written)
followed by V1->V2 hid the V0->V1 change for good. A pending row below the
latest is now served when the NEXT trail row at that time recorded a different
prior value (kind, value, status or the prior's server_ts) - the next change saw
what this one wrote; an orphan of a change that never landed leaves the same
prior for the next row and stays hidden. QuestDB window functions reject
VARCHAR, so the next row's prior comes from a LEFT JOIN on (ts, lead(id)).
The updater's comment and marker-failure log no longer claim modified reads
"still serve it" unconditionally.
Regression tests (QuestDB 10.0.1 throwaway container): marker lost then a
replace / delete at the same time; an orphan followed by a marker-lost retry
then a later change (orphan still dropped); the random marker-mix fuzz now
varies the recorded priors. All four fail on the old filter.
NITs:
- SqlServerHistorianValueWriter: a SqlException raised while the token is
cancelled (SqlClient's "Operation cancelled by user") is rethrown as
OperationCanceledException instead of booked as a failed write. Integration
test holds a TABLOCKX on Sample and cancels mid-INSERT; it returned false before.
- SqlServerHistorianDataUpdater.DeleteRawBatches is an Interlocked counter.
- HistoryReadEvents: an event filter whose SDK validation throws (unknown
operator) failed with the generic BadUnexpectedError; the node manager now
re-throws any failed event-filter validation as BadEventFilterInvalid.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
HistoryReadEvents projected only EventId / SourceName / Time / ReceiveTime /
Message / Severity: a select clause asking for EventType, ConditionName or
SourceNode (what most generic alarm viewers ask for) got nulls, and the
event filter's where-clause (an OfType) was ignored.
- HistoricalEvent gains optional EventTypeName, ConditionName and AlarmId.
SqlServer, QuestDB and Gateway event reads fill them from what their
alarm writers already store (AlarmTypeName / AlarmName / AlarmId; the
gateway event Type and its AlarmName / AlarmId properties).
- The node manager maps the type name to its Part 9 event type (unknown or
non-event => AlarmConditionType), projects ConditionName, and resolves
SourceNode to the materialised condition the alarm id names (the live
events' SourceNode), else null.
- The where-clause is evaluated per event with the SDK content filter
through an IFilterTarget (OfType via the type tree; other operators
against the projected fields), after the page's cursor is taken, so
paging is unchanged.
Tested in the node manager, over a real client session (the Client.CLI has
no `historyread --events` mode yet; filed as a follow-up), and by alarm
writer -> event read round trips on SqlServer and QuestDB.
Rebased from the unpushed remed/lane-H2 branch (originally authored 2026-10-22
in session_01Ub9hGmYLQvn9XETzw7CMsT) onto current master.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
Benchmark wall-clock budget 1 s -> 5 s (measured 5-60 ms; still catches O(N^2)). Host-status wait is already a TCS signal; its 30 s ceiling is raised to 120 s so a starved box cannot time out before the keep-alive reports the loss.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
AbLegacyDriver stamped LastSuccessfulRead (init, read, write), the read's sample
timestamps and host-state change times from raw DateTime.UtcNow; only the backoff
path read the injectable Clock. All of them now read Clock(), so
NativePcccBridgeStatusHealthTests drives a manual clock (init at T0, read at T1)
and asserts LastSuccessfulRead == T1 exactly instead of spin-waiting for the wall
clock to pass the init stamp (#767). Without the driver change the 6 tests fail
(init stamp is wall-clock, not T0).
Directory.Build.props sets TreatWarningsAsErrors for every project since #766, so
the 75 per-csproj copies (64 src + 11 tests) are removed and the props comment
says it is now the only place. tools/mem-probe keeps its explicit opt-out.
Release build: 0 warnings, 0 errors; a planted CS0168 still fails the build in a
src project and in a test project that used to carry its own copy.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
The native-ack path is fire-and-forget after the Part 9 Acknowledge returns, so
a Galaxy ack refused for lack of a credential (#749), or rejected by MXAccess,
left the client with Good and the condition showing Acknowledged while Galaxy
was never acked. Both options from the issue, layered:
1. Synchronous refusal for what the driver can decide without I/O. New optional
IAlarmAcknowledgePreflight (Core.Abstractions); GalaxyDriver refuses an
anonymous session when the driver has no AlarmAck fallback account.
DriverHostActor publishes a condition NodeId -> owning-driver snapshot into
NativeAlarmAckPreflightView (lock-free, volatile); the node manager's new
NativeAlarmAckPreflight seam, consulted after the authz and Primary gates,
returns BadUserAccessDenied with the reason - nothing routed, applied or
reported. A keyed session is not pre-checked (its identity lives in the
Secrets store, which must not be read under the SDK lock).
2. Revert for everything else. RouteAlarmAck carries the condition NodeId;
DriverInstanceActor replies AlarmAckFailed on a refusal, a backend failure,
a driver without an alarm source, or an ack dropped while connecting /
reconnecting. DriverHostActor re-projects the condition unacknowledged via
NativeAlarmProjector.TryRevertAcknowledge, but only while the local ack is
still pending - any driver transition projected since then wins. A disabled
or unspawned driver is not reverted (unchanged).
GatewayGalaxyAlarmAcknowledger now throws DriverOperationRefusedException on a
non-zero native hresult (wrong password, no ack permission) instead of only
warning, so that case reverts too; as a refusal it neither retries nor trips
the breaker.
Tests (each fails with its fix mutated out): projector revert/supersede,
DriverHostActor refused + failed ack -> unacked AlarmStateUpdate (successful
ack control), preflight view snapshot, node-manager preflight unit test,
over-the-wire ConditionMethodEventTests (BadUserAccessDenied, no route, no
event; positive control), Galaxy preflight + hresult refusal.
Docs: Galaxy.md, AlarmTracking.md, ScriptedAlarms.md, .claude/rules/alarms.md.
Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV
mxaccessgw 488649c+ acknowledges by AuthenticateUser as the operator and
writing <alarm>.AckMsg, so AcknowledgeAlarmRequest now needs a real Galaxy
operator_user + operator_domain + operator_password. The OPC UA data plane is
API-key/anonymous and carries no Galaxy password, so the driver resolves the
account per ack:
1. the session's API key's Galaxy ack identity, set on AdminUI /apikeys and
held ONLY in the encrypted Secrets store
(otopcua/apikeys/<keyId>/galaxy-ack-{user,password}); a key with a user but
no password secret is refused, never silently downgraded;
2. else the driver instance's AlarmAck fallback account (DriverConfig
FallbackUser/FallbackDomain/FallbackPasswordSecretRef; secret:/${secret:}/
env:/file:/dev: only — a literal is refused);
3. else DriverOperationRefusedException: nothing is sent, never an empty
operator_user. Refusals are excluded from the resilience retry/breaker.
- Pins ZB.MOM.WW.MxGateway.Client/Contracts 0.2.1 -> 0.3.0 (vendored into
nuget-packages/ from the Gitea feed, per the local-mxgw source mapping).
- The session ApiKeyId (not a secret) rides NativeAlarmAck ->
RouteNativeAlarmAck -> RouteAlarmAck -> AlarmAcknowledgeRequest; the password
is resolved in the driver and never travels through actor messages.
- The OPC UA principal + key id are appended to the AckMsg comment
("[OPC UA: <principal>, key <id>]") and logged, so a shared fallback
account does not lose who acknowledged.
- Password never logged: GalaxyAckCredential.ToString omits it, the
acknowledger scrubs it from every gateway message it echoes and withholds
any exception that echoes it; audit rows carry only the key id.
- AdminUI: /apikeys "Ack identity" editor (set/remove; delete-key clears it),
Galaxy driver form gains the fallback section.
- Synchronous OPC UA status for a refusal is follow-up #760 (the native ack
path is fire-and-forget after the Part 9 method returns).
Tests: request mapping (password populated), password absent from logs and
exception text (mutation-checked), key-mapped / fallback / refused paths at
resolver and driver level, breaker non-trip, key-id propagation through node
manager and DriverHostActor, AdminUI service + form round-trip.
Closes#749
The Modbus lane (#741 U3) gave the driver String[] and BitInRegister[] holding-register array encoders, and
moved address resolution to ModbusTagAddressResolution (keys first, a bit region coerced to Bool, strict key
reads), but DriverArrayWriteSupport still listed both array shapes read-only (follow-up F1), so a client
write was BadNotWritable before it reached the driver.
Modbus row now mirrors the driver end to end:
- Keys shape only (region + address both present). A string-only addressString tag is resolved by the
driver with its own family through a parser Core.Abstractions cannot reference, so it stays read-only
(pinned: the conservative direction).
- Strict reads as the resolver: address an integer 0-65535; a present region / dataType / byteOrder /
stringByteOrder string must name a member.
- Coils: any dataType (coerced to Bool), <= 1968 coils (FC15).
- HoldingRegisters (FC16, never Bool): numeric, String ((stringLength+1)/2 registers, stringLength >= 1),
BitInRegister (1 register, bitIndex 0-15, span ending <= 65535), all within maxRegistersPerWrite.
FC23 is opt-in and never changes the answer.
- Elements convert (ScadaBridge int[] -> Int16 stays writable); only a String / DateTime node over a
numeric or bit element is left out, since such an element never converts.
Tests: the parity test's PendingF1 exemption and its guard are gone; the parity matrix is now region x
element type x node type against the real WriteAsync with a spec-conformant device (FC15 > 1968 refused),
plus shape edges at every cap boundary (numeric, String, BitInRegister, coil), strict key reads, FC23
on/off, string-only, the ScadaBridge int[] shape and U3 wire bytes. Core.Abstractions row tests and a
composer test (String[] and BitInRegister[] writable; over-cap, no-stringLength and string-only read-only).
Docs: table XML doc, Uns.md, modbus-addressing.md, drivers/Modbus.md, data-plane rules.
Refs #741
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The Wave-2 lanes gave AbCip, AbLegacy and TwinCAT real whole-array write
encoders (review 07 U10/U9), but the capability table still listed all three
as scalar-only, so every array node materialised read-only and a client write
was BadNotWritable before it reached the driver.
Rows added, each listing only what WriteAsync encodes end to end:
- AbCip: SInt/USInt/Int/UInt/DInt/UDInt/Real/Dt/LInt/ULInt/LReal, declared
length, not a safetyTag. BOOL/STRING/Structure arrays stay read-only.
- AbLegacy: Int/AnalogInt on N/B/A, Long on L, Float on F (file-width match),
plain <file><n>:<word> address, <= 256 elements; not /bit, .sub, indirect,
S:/I:/O:, nor native PLC-5.
- TwinCAT: every codec element type, declared length; not a BOOL bit-in-word
path.
Elements convert in all three encoders, so the rows never key on an exact
node DataType (ScadaBridge int[] -> Int16 stays writable); they leave out
only node types whose elements never convert (DateTime/String over numeric,
Boolean on AbLegacy, STRING under DateTime on TwinCAT: culture-dependent).
Tests: driver-level parity for each driver (every element type x node type
plus shape cases, against the real write path), the AbLegacy parity tests
un-skipped, Core.Abstractions row tests, a composer test (three drivers
writable, refused shapes read-only), and the existing AbCip negative cases
moved to a BOOL array. Docs: Uns.md table, per-driver docs, data-plane rules.
Refs #737#738#739
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
N3 (critical): the node manager carries only NATIVE conditions' state over a
full rebuild; scripted conditions come back fresh and rely on the engine's
re-assert, which since the off-thread plan is handled in the window and lands
on the discarded instance. On a rebuild the condition replay scope is now
AddedConditions plus every enabled scripted condition in the new composition
(native kept conditions stay carried-over, not replayed). The test sink again
models the real rebuild: it clears non-native condition state only.
N4: a surgical in-place shape edit (DataType / IsArray / ArrayLength) resets the
raw node and its UNS mirror to BadWaitingForInitialData without ensuring them,
so the window's value was not replayed and the two NodeIds disagreed. The
applier records re-shaped nodes (MaterialisationRecording.Reshaped); their
deferred value replays only once fitted to the new shape (VariableShape.TryFit:
as is, or converted invariantly; never an old-typed value - WriteValues assigns
the value verbatim with no type check), otherwise it is dropped at Debug and the
node keeps waiting. Status-only entries are not replayed onto a re-shaped node.
Minors: m1/m2 documented on _replayScope (degraded-path only); m3 corrected the
false carry-over premise in the scope doc (the ScriptedAlarmHostActor ordering
remark is outside this lane - F8); m4 the replay Debug log reports alarm-state
and alarm-quality counts again.
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
N1: a full alarm-state snapshot also sets the condition's Quality, so it now
supersedes an earlier quality-only write for the same id in the deferred
buffer; a quality arriving after the state is kept and replays after it.
N2 / M1: the replay is not inert on a node the apply KEPT - an SDK-applied
Acknowledge/Confirm/Shelve or an optimistic client Write in the window would
be rolled back (with spurious Part 9 events). The replay is now scoped to
nodes the apply created: variables the applier ensured during this apply
(AddressSpaceApplier.BeginRecordingEnsuredVariables) or every variable after a
full rebuild; conditions in the applied composition but not the previous one
(also on a rebuild, which carries a kept condition's state over). Option (b),
a sink that reports whether a live write landed, needs the Commons
DeferredAddressSpaceSink seam outside this lane (follow-up F7).
M2: the benchmark remark now matches its Category=Benchmark trait.
M3: TestHooks and the seam-carrying constructor are internal; PropsForTests
produces the actor through an IIndirectActorProducer.
M4: pinned replay-nothing-and-drop-the-buffer after a failed plan.
M5: PreRestart answers every queued rebuild requester (degraded).
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
C1 (critical): B6 computes the plan off the pinned thread while B1 made writes
non-creating, so a value/status/alarm write for a node the in-flight plan adds
was skipped, and change-driven drivers never republish it (BadWaitingForInitialData
indefinitely; on a full rebuild every re-subscribe value landed on discarded
nodes). OpcUaPublishActor now applies writes immediately AND remembers the
latest per (realm, id) while a plan is in flight or queued (value+status
compose; alarm state/quality coalesce to the latest, which the node's delta
gate makes inert on an existing condition), replays it after each apply before
answering, and drops it once the pipeline is idle. Bounded (250k keys by
default; beyond it new keys are not held, logged once).
I2: the data-plane-authority parse runs in its own try inside the plan task; a
throw is logged and never degrades the address-space apply.
I3: the SDK's DeleteMonitoredItems releases an item's component-cache
reference on a Good delete; a tombstoned item's reference was already
released, so the delete took one from the re-created node's entry. Released
items are remembered and the orphan delete takes one back first. Event items
never held a reference (EventManager + SubscribeToEvents), so the three
event-item releases are gone.
I4: direct tests for the HistoryUpdate override's per-request updater check
and the HistoryReadEvents override's source guard (both reached by restoring
the bit after a reset - the B7 race window); fixed the false remark.
Minors: the MaterialiseAlarmCondition kind-swap tombstones the dropped
instance; a bytes-less rebuild arriving mid-plan queues behind it (replies in
order); a PlanReady/PlanFailed from a previous actor incarnation is ignored and
its requester answered degraded; the tie-cluster end bound of an uncapped read
fits against the backend ceiling (defence in depth - guarded by a sweep test);
the lock-time benchmark stays in the default run, headroom noted.
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
B6 moved the address-space-rebuild span into ApplyPlan, so the two rebuilds
that complete synchronously on the actor thread (no applier -> raw-sink
rebuild; no artifact bytes -> held) stopped emitting it
(OtOpcUaTelemetryHookTests.RebuildAddressSpace_starts_an_address_space_rebuild_span).
Both now open their own span.
B7 (late HistorianDataSource / HistorianDataUpdater re-evaluate the
capability bits) reversed three tests that pinned the old "fixed at
materialisation" contract. Re-expressed against the new one, end to end:
- events: a folder promoted without a source is unsupported, then served
once a source is wired (source reached);
- in-process HistoryUpdate: clearing the updater retracts HistoryWrite and
the old updater is never reached;
- over the wire: a late updater lights HistoryWrite and an insert reaches
it; a reset retracts the bit and the insert is BadHistoryOperationUnsupported.
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Lane H1 follow-through: SyncAdvertisedAggregates now also calls
HistoryPaging.AdvertiseMaxReturnDataValues(Server, source
.MaxRawValuesPerRead) so HistoryServerCapabilities/MaxReturnDataValues
(i=11273) follows the historian source on every assignment (the Null
default advertises 0 = no limit); a false return logs a Warning (the cap
is enforced by HistoryRead regardless).
Test: NodeManagerMaxReturnDataValuesSyncTests reads i=11273 back through
the SDK master node manager after setting a capped source (7000, 250)
and after resetting to the Null source (0).
Refs #732#736
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
H1 fuzz finding (routed to Lane B because AppendEndBoundAsync lives in
OtOpcUaNodeManager.cs): a bounded HistoryRead-Raw page sized BELOW
NumValuesPerNode by the backend's row ceiling (RawPageLimit.Plan) took
the end bound whenever its data fitted, because the fit was judged
against NumValuesPerNode instead of the page's effective cap — a full
ceiling-sized page came back with cap + 1 values. The bound is now
fitted against the page allowance (the ceiling-adjusted data cap plus
the start bound a first page already carries); when the page is full
the bound goes to the next page through the continuation point.
Test: NodeManagerHistoryReadBoundsTests
.A_full_ceiling_sized_page_carries_the_end_bound_on_the_next_page
(ceiling 4 ⇒ 3 data rows; first page = start bound + 3 with a CP, second
page = the end bound alone; control: no ceiling ⇒ one page of 5).
Refs #732#736
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Review 01 U3. EnsureFolderIsEventNotifier granted the EventNotifiers
.HistoryRead bit (and registered the event-history source) only when a
real HistorianDataSource was wired at promotion time, and EnsureVariable
composed the AccessLevels.HistoryWrite bit from whether a
HistorianDataUpdater existed at materialise time — so an assignment
after the first apply left existing folders without event history and
existing nodes refused by the SDK's HistoryUpdate admission until a full
rebuild, and a source reset to Null never took the bit back.
The two setters now re-evaluate: HistorianDataSource walks every promoted
folder (bit + source registration, both directions), HistorianDataUpdater
recomposes the AccessLevel of every historized variable (Writable read
back off the node's own CurrentWrite bit). The Host's boot ordering is no
longer load-bearing.
Tests: NodeManagerCapabilityRefreshTests (late source lights the bit on
an already-promoted folder and a reset takes it back, live-event
subscription untouched; late updater grants HistoryWrite to historized
nodes only and a reset takes it back, authored Writable preserved).
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Review 01 U2. The composer never emits EquipmentTags in v3 (device values
live in the Raw subtree and reach equipment as UnsReferenceVariables), so
the equipment-tag plan diff sets (Added/Removed/ChangedEquipmentTags,
EquipmentTagDelta), the planner's diff over them, the classifier's
TagDeltaIsSurgicalEligible and its Rebuild clause, the applier's
MaterialiseEquipmentTags pass, surgical ChangedEquipmentTags loop,
removed-tag teardown loop and add-announcement term, and the publish
actor's equipment-tag pass were unreachable, maintained dead weight.
IsEmpty keeps its semantics (an otherwise-empty plan is still Empty).
Kept, deliberately: AddressSpaceComposition.EquipmentTags and the
EquipmentTagPlan record — Runtime's DeploymentArtifact (outside this
lane) still assigns the empty set and the artifact parity tests assert
it; follow-up F2 removes both together.
Tests: the equipment-tag vehicle tests for PureRemove / AddRemoveMix /
the rebuild ratchet are re-expressed over raw tags (same applier paths);
the equipment-tag surgical and classifier tests are deleted (their raw /
UNS-reference equivalents already exist); SubscriptionSurvivalTests now
materialises its tags through MaterialiseRawSubtree.
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Review 01 L3/H5/H4/P6/P4.
- L3/H5 OnEquipmentTagWrite's dispatch is factored into DispatchDeviceWrite:
a gateway that throws synchronously is refused inline with
BadDeviceFailure (it used to escape the handler under the SDK Lock), and
a faulted/cancelled dispatch task has its exception observed and logged
and carries the real reason into the revert/audit event instead of
surfacing as UnobservedTaskException at finalization.
- H4 every timestamp the manager stamps (condition quality, timed-shelve
expiry, method-transition events, carry-over, revert, revalidation)
reads the injected Clock; RemainingShelveMs takes the clock reading.
- P6 OpcUa:Quotas:* (OpcUaQuotaOptions on OpcUaApplicationHostOptions):
Min/MaxRequestThreadCount, MaxQueuedRequestCount with the former
hard-coded 5/100/200 as defaults, and the TransportQuotas members as
nullable overrides over the SDK defaults.
- P4 the SDK line exposes IHistoryReadAsyncNodeManager but CustomNodeManager2
offers no async virtual to override (adopting it means re-implementing
the base dispatcher), so the lever is the admission cap:
MaxConcurrentHistoryReadBatches defaults to clamp(2*ProcessorCount, 4, 16).
The Host still overrides it from ServerHistorianOptions (default 16) —
follow-up F3.
Tests: NodeManagerWriteDispatchTests (sync throw refused; faulted task
reverts; success leaves the value; clock-stamped revert),
OpcUaApplicationHostQuotaTests.
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Review 01 P1/P2/P3/P5 — work done under the manager-wide SDK Lock on the
pinned publish thread, where every Read/Browse/Publish waits.
- P1 RebuildAddressSpace detaches the shared root's top-level children
once instead of every leaf from its parent (NodeState.RemoveChild is a
linear scan + RemoveAt, ~k²/2 compares per k-child folder); root
notifiers are cleared as a list instead of one O(F) RemoveRootNotifier
each; ForgetPredefinedSubtree / TombstoneMonitoredSubtree are iterative
with one child buffer.
- P2 RemoveEquipmentSubtree walks a parent→children index maintained by
EnsureFolder/EnsureVariable/MaterialiseAlarmCondition (keyed by the
parent id the caller named) instead of prefix-scanning six maps per
removed equipment; a surviving condition's wiring into a removed folder
is found through the folder's own notifier list; descendants are not
detached from a folder that is itself being discarded.
- P3 LowerAllEventsMonitoring(condition) replaces the full O(C) reconcile
on RemoveAlarmConditionNode / RemoveEquipmentSubtree / the kind-swap.
- P5 AlarmNotifierWiring.Folders is a reference-identity HashSet and an
already-wired pair skips the SDK's linear AddNotifier dedup.
Benchmark (NodeManagerLockTimeBenchmarkTests, M-series Mac under load,
three runs each): rebuild of a 10k-tag folder 75–150 ms → 5–6 ms;
removing 200 equipment (250 signals each) from a 50k-node space
471–681 ms → 58–64 ms; all-events bookkeeping visits for 10 condition
removals over 50 conditions 900 → 10. The two timing legs assert < 1 s;
the P3 leg asserts the visit count (machine-independent).
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Review 01 L2. The SDK keys its fan-out lists by NodeId and its component
cache hands back the CACHED NodeState for a NodeId it already holds, so a
rebuild or a surgical remove+re-add left a NEW item on the same NodeId —
from any session — bound to the DEAD node: created Good, fed its stale
initial value, never again, until every session had deleted its items.
Every removal path (RebuildAddressSpace, RemoveVariableNode,
RemoveAlarmConditionNode incl. the condition's instance-scoped children,
RemoveEquipmentSubtree) now queues BadNodeIdUnknown to the node's data
items, drops the MonitoredNodes entry, releases each item's component
cache reference and forgets the _liveItems/_eventItems entries. Items
the ACL pass stood down (on no fan-out list) are swept by NodeId. A later
client delete of an orphaned item is completed by the DeleteMonitoredItem
override / the SubscribeToEvents unsubscribe arm instead of answering
BadMonitoredItemIdInvalid or NRE-ing in MonitoredNode2.Remove. The
revalidation restore re-resolves the node by NodeId and no longer
re-attaches to Handle.Node blindly. Timestamps now come from the injected
Clock (UtcNow).
Tests: StaleMonitoredNodeWireTests (real SDK, three legs: rebuild,
surgical remove+re-add, second session; each keeps the old item alive,
proves the new item receives a value written after its creation, the old
one gets BadNodeIdUnknown, and its delete succeeds — all three fail with
the tombstone neutered), NodeManagerStaleMonitoredNodeTests (stood-down
item swept on surgical remove and rebuild; sibling removal leaves it
restorable).
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Review 01 L1/L4/H2. WriteValue/WriteValues are non-creating (an unknown
id is skipped, counted and Debug-logged, as WriteStatuses always was), so
a driver publish racing a surgical remove can no longer resurrect the id
as a flat BaseDataType phantom under the shared root — nor hand a later
re-add of the same tag the phantom's parent/DataType/writability through
EnsureVariable's early return. WriteAlarmCondition for an unmaterialised
id is a logged no-op instead of minting the legacy bool[2] placeholder
(AddPredefinedNode is AddOrUpdate, so it replaced the SDK lookup entry of
any folder sharing the id). The applier no longer writes a condition
state against every removed equipment id (a folder, never a condition).
EnsureFolder/EnsureVariable/MaterialiseAlarmCondition refuse an id already
held by another node kind in the same realm. CreateVariable is deleted.
Tests: NodeManagerPhantomNodeTests (publish racing a remove leaves no
node; the re-add gets the authored shape and is live; unknown alarm id
is a no-op; the guard in all three directions, realm-qualified); the
sink / batch-write / write-statuses tests now declare their variables.
Refs #732
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- GatewayHistorianDataSource.ReadRawWindowAsync keeps the cap + 1 probe budget
(budgetCap) separate from the client's cap: a strict-ceiling or crowding
refusal now reports Limit = the server cap that applied and advises setting
NumValuesPerNode at or below it (it said "lower NumValuesPerNode" with
Limit = cap + 1 to a client that sent 0).
- Fuzz: the oracle-source over-cap assertion was vacuous (cap 0 never counts);
it now checks no reverse window reached the forward-only shape. Class docs
name the post-floor cap set (3 ... 1000) and distinguish NumValuesPerNode
1 / 2 from the smallest server cap.
- HistoryPaging.IsFullPage doc says it judges the REQUEST cap (page + ties +
lookahead), which is finite under any limit; AdvertiseMaxReturnDataValues
swallows any exception (returns false for the caller to log), as its doc
promised.
Refs #736
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- NodeManagerHistoryReadRawCapFuzzTests: 10k seeds (env OTOPCUA_PAGING_FUZZ_ITERATIONS)
through a real SDK node manager against wire-faithful fakes of SqlServer / QuestDB /
Gateway that enforce the backend half of the cap; random series with ties, random
windows (empty, on/past a tie, between samples, forward/reverse/end-only, bounds),
N in {0,1,2,small,huge}, cap in {off,3..1000}, tie budget {3,8,65536}. Pages must
concatenate to the cap-off one-shot answer (checked against the series), terminate
within values + 2 (+ one empty step per oversized boundary cluster), never exceed
the cap, never ask the backend for 0 / over-cap; refusals only past the budget.
- HistoryPaging.AdvertiseMaxReturnDataValues writes HistoryServerCapabilities/
MaxReturnDataValues (i=11273), read back through the SDK master node manager.
- MinMaxRawValuesPerRead raised to 3: at 2 a resumed page has no room (value + its
boundary tie + lookahead) and every value cost an empty page + CP (seed 15).
- RawPageLimit rows pin the default-cap arithmetic (9 999 first page, 9 998 resumed).
Fuzz findings recorded as follow-ups: a bounded read whose data fits one page
returns cap + 1 (end bound fitted against N, not the cap; node manager, F2);
the Gateway walk-back's cost on a sparse series under a tiny tie budget is
pre-existing (seed 320: 82 reads uncapped, 93 capped).
Refs #736
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Item 1: IHistorianDataSource.MaxRawValuesPerRead (Gateway: learned ceiling - 16 = L; 0 elsewhere, guard-tested).
RawPageLimit plans every raw page on top of #692's lookahead: with L > 0 the read (page P + boundary ties k
+ lookahead) is <= L, P = min(N, L - k - 1) (uncapped: L - k - 1); no room => straight to the tie-cluster
over-fetch (never an uncapped or underflowing request). One re-plan + re-read when the read taught a new
limit or was refused for crowding pre-window rows (HistoryReadLimitExceededException.PreWindowRows): the
page shrinks by them. Over-fetch bound min(MaxTieClusterOverfetch, L - 1); slice size never 0.
Review fixes: F1 a window starting at/before the epoch counts AVEVA's synthetic row as pre-window (a
ceiling-truncated read was reported complete); F2 the page path disables ReverseRawReader's own re-run
(cold end-only ceiling 20/N=3000 cost 3,023 reads, now <= 40); F3 crowded look-back re-read around, not
refused; F4 the #687 crowding re-read room at least doubles (up to 8 re-reads), so a cluster of hundreds
just before a resume point no longer fails on a default gateway; a crowding refusal that leaves no room
is BadTooManyOperations, never a tick-by-tick cursor.
Item 2: ServerHistorian:EventReadHorizon (default 00:00:00 = exact, unchanged). When set, every fresh
HistoryReadEvents window's earlier bound is raised to now - horizon (both directions, all backends).
Lossy by design; documented.
Item 3 (native TimeAverage/Total re-check) stays blocked on historiangw#27.
F1 (pre-existing): SliceTieCluster continued a forward read while next <= endUtc, but the
forward window is end-exclusive; an oversized cluster one step below EndTime resumed AT
EndTime, the resume became a start == end point read and returned the EndTime samples as
data (microsecond: EndTime 8.1 us truncated to 8.0). Continue only while next is strictly
before EndTime truncated to the backend resolution.
Tests: A_tie_cluster_just_below_the_end_time_never_resumes_onto_the_end_time (native +
microsecond; both fail on the old rule) and
A_cursor_inside_a_tie_cluster_over_fetches_without_a_normal_read_first (pins the
WithinTieCluster shortcut by per-page read count; fails with the shortcut disabled).
Docs: the oversized-cluster exception to 'every page full / no lookahead => last' (F2),
and the N + skip + 1 strict-ceiling edge note referencing #694 (F4).
Review follow-up: on the Gateway every value is a Double (AVEVA stores a Boolean
as a 0/1 Int2), so the raw-evaluated AtTime interpolated a Boolean node to 0.5.
- IHistorianDataSource gains a default-interface overload
ReadAtTimeAsync(fullReference, timestamps, TypeCode? valueType, ct) that
forwards to the existing one (SqlServer/QuestDB/Null/fakes unchanged).
- OtOpcUaNodeManager.HistoryReadAtTime passes the node's scalar DataType
(AtTimeValueType: built-in type -> TypeCode; arrays/other types => null);
Raw and UNS NodeIds alike.
- GatewayAtTimeReader converts each raw row to the hinted kind BEFORE the shared
AtTimeEvaluator runs (Boolean = value != 0 => stepped; integers => Int64 =>
interpolated, rounded half away from zero, returned as the node's integer
type; floats Double), so the evaluator's own type branch decides.
- AtTimeEvaluator boxed every interpolated Int64 as a Double (`cond ? (long)x :
v` widens to double); its "keeping its type" test passed because Shouldly
ShouldBe compares numbers across types. Fixed (affects SqlServer/QuestDB too)
and pinned with ShouldBeOfType.
A resumed raw page asked the backend for only N rows from the boundary timestamp
(inclusive), then trimmed the boundary ties already emitted: it netted N - skip values,
and at N = 1 nothing — the empty "full" page was taken for a stalled tie cluster and a
MaxTieClusterOverfetch + 1 (65,537-row) over-fetch stepped one tick, returning an EMPTY
page with a continuation point between every value (8 values = 17 pages / 25 reads),
forward and reverse, with or without ReturnBounds, on every backend.
Each page's read now asks for N + boundarySkip + 1 rows: the ties it trims, the page and
one lookahead row. The lookahead row is the 'more data' signal, so an exact multiple ends
on its last full page (no trailing empty page). The over-fetch fires only when every row
past the trimmed ties is still on the boundary (a genuine cluster larger than the page);
a cursor left inside a cluster is flagged WithinTieCluster and over-fetches directly.
A page that never leaves the boundary adds the prior skip to its cursor.
Low items: BoundedReadHasAnythingAsync carries its walk hint and, when there is no data,
the end bound it found (HistoryEndBound.Resolved + DataExhausted), so the next page reads
nothing; HistoryBounds.IsPlaceholder is an explicit marker (ConditionalWeakTable over the
NotFound instances) instead of status+null shape, so a stored Bad_BoundNotFound row is data.
Tests: NodeManagerHistoryReadPageFillTests (N=1 every shape/direction/bounds, tie clusters
straddling boundaries at N=1/2/3, exact-multiple, carried end bound + walk hint, stored
placeholder-shaped row). Existing paging tests updated where they encoded the short pages.
1. The at-or-after bound read on a forward-only backend (Gateway) ends at
now + ReverseRawReader.LookAhead instead of DateTime.MaxValue (the gateway
only checks start <= end, so a year-9999 end reached AVEVA); a bound beyond
the horizon is not found without a read. SqlServer/QuestDB keep MaxValue:
they bind it safely and may hold samples further ahead. The Gateway fake now
refuses far-future ends, so the missing-end-bound test is no longer vacuous.
3. No data and no bound found (whole read empty) => GoodNoData, no values —
the pre-bounds answer (6.5.3.2: success only if one or both bounds exist).
A one-value page whose start bound is missing first checks there is data or
a found end bound.
4. ServerHistorian:HonorReturnBounds (default true); false ignores ReturnBounds
as before #691, including the modified-read Bad_InvalidArgument.
The node manager composes the start/end bounds from each backend's existing raw
reads (one authority, HistoryBounds): at-or-before = a capped reverse read
(native, or ReverseRawReader's walk-back on a forward-only backend — never
AVEVA's synthetic boundary row), at-or-after = a capped forward read. Bounds
count toward NumValuesPerNode, the start bound leads the first page, the end
bound ends the last (carried in the continuation point). Missing bound =
Bad_BoundNotFound (null, at the bound time; one-sided reads ±1 s from the
previous value, Table 1 footnotes a/b). A modified read with ReturnBounds is
Bad_InvalidArgument (Part 11 6.5.3.3).
The .NET SDK's ReadRawModifiedDetails defaults ReturnBounds to TRUE; the
existing tests built details without setting it and now set it false.
Oracle: every row of Part 11 Table 1 over three backend shapes, plus paging
concatenation, ties, Bad bounds, microsecond truncation; conformance pins the
lookup primitives on SqlServer + QuestDB.
A tie cluster's truncated read left the width at one tick; once the gap
above it was under two milliseconds nothing floored it, so segHi
stepped down one tick per read: 30 rows at one ms 0.3 s below hi with a
carried 1 s width took 10,006 identical reads (S8). (a) The top-anchored
width is floored at the granularity, Min kept; (b) a segment no wider
than the granularity is the unit that escalates to
MaxTieClusterOverfetch and then refuses - per millisecond, exact for the
ms-resolution Gateway (documented). Both are needed: the floor alone
loops on crowded reads with a small tie budget. S8 is a repo test
(4 shapes, reads <= 30, 0 repeated requests); random-data tests now
measure the refusal unit as rows in any 1 ms window. Reviewer harness,
5,000 seeds x default/MSONLY/CROWD: worst page 70 reads (default/MSONLY),
63 (CROWD) at M=65536, p99.9 27-37; no page over 500 reads at any tie
budget.
Bisecting the gap above a truncated read halved below the 1 ms request
granularity (requests are ms-rounded, so it re-sent identical requests)
and bisected the WHOLE gap even when the burst ended right above the
row it saw: a 1,026-row burst ending 100 days ago took 46 reads. After
an empty top-anchored segment the walk now gallops up from the known
row ((known + d, top], d x4, following each truncated read), then
bisects the bracket, with halving floored at the granularity (the Min
kept; a crowded gallop falls back to top-anchored halving). Burst: 18
reads, 0 repeated requests; dense tag stopped 2 days ago: 26 (gallop
alone crawled: 94). Walk fuzzed at 4,000 seeds locally (400 in-tree).
Docs: measured counts; CLAUDE.md gateway prerequisites list
RuntimeDb:EventReadMaxRows.
A reverse or end-only capped events read on a backend without
SupportsReverseEvents (Gateway, Null) was refused
BadHistoryOperationUnsupported (explicit reverse) or answered with the
OLDEST N ascending (end-only, #675). It is now served newest-first by
the same ReverseWalk as raw reads: page-sized ms-aligned segments,
never answered from a read the source flagged truncated, ties sorted by
EventId (the gateway orders by event time only - the fake shuffles
them per call), one extra event walked to signal 'more', uncapped = one
forward read reversed, and the width + proven-empty stretch carried in
the events cursor. The oversized-cluster over-fetch keeps exactly the
events at T in the same order. A tie cluster over the budget is
BadTooManyOperations. Docs note the gateway's EventReadMaxRows must be
unlimited (it truncates silently).
The last pages of a deep reverse page-through on a forward-only backend
each walked down to 1970 one x4 growth step at a time (~14-16 reads,
~2.5 s live). After two empty segments the walk now asks once per page
for the whole remainder at the page budget (a truncated answer is
discarded and the walk carries on), and the stretch the page's last
complete read proved empty rides the continuation point
(HistoryContinuationState.EmptyFromTicks/EmptyToTicks, session-held, so
no versioning) so the next page steps over it. Fuzz now pages with the
walk's own hints and with true random empty stretches.
ReverseRawReader aimed every segment at half of MaxTieClusterOverfetch
(~32k rows by default) and carried that width page to page, so a paged
'latest N' read on a dense Gateway tag fetched ~32,770 rows per page
(2-5 s live) to return N. The walk now lives in a generic ReverseWalk
core: segments are aimed at ~N+8 rows (page budget 2N+16), a discovery
segment (first without a hint / after an empty one) may read 1,024, and
the tie budget is only the ceiling a single-timestamp cluster escalates
to. A truncated read never re-enters the stretch below its newest row
until the gap above it is bisected. Tests: per-page row counts on a
dense tag (node manager + walk), randomized exactness fuzz (walk and
node-manager paging), dense-tag-below-a-long-gap read bound.
1. Uncapped reverse raw read on a forward-only backend = ONE forward read of the
window, filtered and reversed in memory (was a segment-by-segment walk).
2. Segment width is aimed at ~half the row budget from the density just observed
(overflow: from the truncated read's own sample span) and carried page to page
in HistoryContinuationState.SegmentWidthTicks.
3. The walk starts no later than now + 1 day and never reads before the epoch.
4. Request bounds are rounded outward to the millisecond, so a row AVEVA
fabricates at a rounded-up query start cannot land inside the segment.
5. Budget M = max(1, MaxTieClusterOverfetch); reads ask M + 2 rows, completeness is
'short OR last row past the segment', and a read crowded by rows before the
segment is re-read once (<= 2M + 4). A cluster of exactly M, neighbouring
clusters each within M, and M <= 1 are all served; only one timestamp over M
is refused. ServerHistorianOptions text updated.
6. The raw and events tie-path oversized-cluster refusal is BadTooManyOperations
(was BadHistoryOperationUnsupported), consistent with the walk.
7. GatewayHistorianDataSource.ReadRawAsync: a caller cancellation is not a backend
failure; NumValuesPerNode = 0 is sent as int.MaxValue (the client package and
the gateway reject max_values <= 0, so every uncapped Gateway raw read failed);
a MaxValuesCeiling refusal is a HistoryReadLimitExceededException.
8. Tests: upper-edge filter, epoch floor, look-ahead, ms alignment, exact/adjacent
clusters, minimal budget, carried density, and the walk over the real Gateway
source + FakeHistorianGatewayClient (OpcUaServer internals visible to it).
#669 turned an end-only capped HistoryRead-Raw into a reverse read (start > end)
for every backend; the gateway rejects start > end (InvalidArgument), so on the
default Gateway backend every end-only capped and every explicit reverse raw read
failed the node and was counted as a backend health failure.
- IHistorianDataSource.SupportsReverseRaw (default false; SqlServer/QuestDB/Null
true, Gateway false; guard test pins that every production backend declares it).
- OtOpcUaNodeManager sends a reverse window only to a backend that declares it;
otherwise ReverseRawReader walks (lo, hi] back in bounded forward segment reads
(1 min, x4 while sparse; each <= MaxTieClusterOverfetch + 1 rows; a full read is
re-read narrower, never answered truncated; filtered on sample timestamps, which
also drops AVEVA's fabricated query-start rows; stops at the Unix epoch). A
one-timestamp tie cluster over the bound => BadTooManyOperations.
- ServeRawPagedAsync maps HistoryReadLimitExceededException to BadTooManyOperations.
- GatewayHistorianDataSource.ReadRawAsync refuses start > end unsent, without
recording a health failure.
ServerCapabilities/AggregateFunctions (i=2997) is Part 5's list of every
aggregate the server supports, and live AggregateFilter subscriptions still
compute all 37 SDK aggregates, so it keeps the SDK's full set; only i=11201
follows the backend. A null SupportedAggregates is treated as the interface
default (all) by both the advertisement and HistoryRead-Processed (it used to
NRE inside the batch). Docs: i=2997/i=11201 split and the boot window.
Co-Authored-By: Claude <noreply@anthropic.com>
IHistorianDataSource.SupportedAggregates (default: every HistoryAggregateType)
is the one authority for both ServerCapabilities/AggregateFunctions (i=2997) and
HistoryServerCapabilities/AggregateFunctions (i=11201) and for the up-front
BadAggregateNotSupported in HistoryRead-Processed. Gateway declares
AggregateModeMapper.Served (4), SqlServer/QuestDB AggregateFamilies.Served (17),
Null none. Only the folder references are trimmed; the SDK's calculator
factories stay so live AggregateFilter subscriptions keep working.
Co-Authored-By: Claude <noreply@anthropic.com>
Joseph Doherty