Coordinator ruling: a delivered chunk proves the Agent is alive (the analogue of
OpcUaClient's good keep-alive), so TryMarkStreamDelivered - generation-checked
under _subscriptionLock - also clears _writeDegraded. The degradation no longer
sticks forever on a subscription-only driver. Test: unanswered write -> Degraded,
stays Degraded with nothing else happening (positive control), the next chunk ->
Healthy; red with the clear removed. Round-2 doc text corrected.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- N-4: the /sample Unsupported latch is set inside TryDegradeStream's
generation-checked critical section (latchUnsupported), so a superseded pump can
no longer latch it on the session that replaced it.
- N-5: a read served from the live pump index clears only the read degradation -
it is stream evidence, like a chunk, which by the I-3 ruling does not clear an
unanswered write. The 'cannot stick' wording in the XML doc and
docs/drivers/MTConnect.md now states the latch-until-next-answered-/current-read-
or-write behaviour, incl. the subscription-only consequence. Tested (chunk +
live-index read keep Degraded; an answered write clears it).
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- N-1 (Important): m3's BadNodeIdUnknown throw for an ack with no resolvable
condition counted toward the host's AlarmAcknowledge breaker (threshold 5), so
re-acks of one unresolvable condition blocked every genuine ack. It now logs a
Warning and returns; only a no-reply transport failure throws. Test through the
real CapabilityInvoker: six unresolvable acks, then a genuine ack that is still
delivered (red with the throw restored).
- N-2: ShutdownAsync holds _subscriptionGate for the whole teardown (in-flight
subscribes finish first, bounded by the budget), so a subscribe arriving during
the delete phase finds the session released instead of creating a subscription
on the closing session. Red-tested.
- N-3: the shutdown-side cert-validation detach is generation-checked under
_probeLock with the map / endpoint clear.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- m1: the form hint now says ProbeTimeoutSeconds can only shorten Test Connect's
10 s budget (passing the configured value as the button's budget is shared code -
follow-up).
- m2: a superseded init no longer overwrites _options (generation-checked under
_probeLock) or the cert-validation handler (handed back and published with the
session; detached when not installed; a previous one detached on replace).
- m3: an ack none of whose condition ids resolves throws BadNodeIdUnknown + Warning.
- m4: an ack with no reply (transport failure / non-caller timeout) degrades.
- m5: no session before the gate is the same ServiceResultException(BadNotConnected).
- m6: the caller-cancelled-write deviation is documented.
- m7: a reply during a published outage advances the stamp but does not write
Healthy; the keep-alive recovery ends the outage.
- m8: test for a non-caller OperationCanceledException write -> Degraded.
- m9: the browse session owns and disposes the user certificate it loaded
(OpcUaUserIdentityFactory.LoadCertificate; a key-less cert is disposed before the
throw); AdminUI-host certificate path documented.
- m10: a subscribe overtaken by a shutdown / re-init removes its half-built
subscription instead of registering it (data and alarm).
- Sync Dispose drains each gate for 500 ms (was 2 s).
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
2: only a PUBACKed (QoS >= 1) publish advances LastSuccessfulRead; a QoS 0 Sparkplug
command completing means only that the bytes left the local socket.
3: a no-verdict SUBSCRIBE caused by the caller's own cancellation now throws
OperationCanceledException (refs still Bad + pending), so the breaker no longer
counts a pipeline timeout as a failure.
4: the rebirth wait is the window capped at 15 s, below the AdminUI 20 s per-call
timeout.
6: a 10-topic test pins the 8-in-flight publish bound. 7: partial-level wildcards and
$share filters are pinned never to confirm. 8: the intended init-vs-shutdown
supersede is commented at the re-check.
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- m6: a /current read that sees a new Agent instanceId now also publishes the new
process's cursor (generation-checked, captured with the read's session), so a
later subscription does not open at the dead instance's sequence and flicker
through a clear + re-baseline. Pinned in the T-2 test.
- Documented (code + docs/drivers/MTConnect.md): m1 the bounded stale-read window of
a stream stalled without failing; m4 a refused multi-pair PUT's retry may re-send
an already-applied pair (harmless); m5 a non-Agent 200 now reads reachable to the
ping; m8 a refused raised count takes one transient blip before the step-back.
- m10 confirmed: no host key is persisted (DriverHostStatus dropped, #521) - the
redacted HostName only changes an in-memory /hosts key for a query-bearing URI.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
One stray duplicate row used to fail every tag in a key-value group or none of them,
depending on whether an unrelated key was absent. Duplicated keys are now refused
per key: only that key's tags publish BadConfigurationError and the rest of the group
reads normally. The key-value cap reads up to 2x the distinct keys (+1 to detect
excess), so a few stray duplicates are attributed per key and only a gross excess
(history/events table) fails the whole group. Duplicated-key refusals count as
answered refusals for health. The row-cap, duplicate-key and multi-row warnings now
use the per-plan limiter (m-3); the reader-wide limiter is removed. Unit tests pin
both the all-keys-present and absent-key cases with Good controls; live SQL Server
gate added (23/23).
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Coordinator ruling on F3: a write with no reply (transport failure / deadline)
sets Degraded with the stamp preserved and a type-only LastError, through a new
_writeDegraded flag folded into PublishHealthy. One success writer is shared:
RecordReadAnswered (a /current answer, an Agent refusal, a live-index read)
clears the read AND write degradations; RecordWriteAnswered clears the write
one; neither clears the other path's stream degradation. Tests both directions
(answered write clears it, good read clears it) plus the converse (an answered
write does not clear a failed read); red with each half removed.
m7: a good /current resets the refusal-dedup latch, so a refusal recurring after a
recovery is logged again (tested).
Refs #743
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
Health contract: an MTConnectError document (NO_DEVICE, INVALID_REQUEST, ...) is the
Agent answering. The parsers now mark such an InvalidDataException
(MTConnectXml.IsAgentRefusal - the type is sealed, so a Data marker, every other
handler unchanged); ReadAsync reports the batch Bad but does not degrade the read
path, advances LastSuccessfulRead, clears a prior read degradation and warns once
per distinct refusal. An unparseable answer stays a comms fault.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The driver logged and surfaced the raw AgentUri (which may carry user:pass@) as
HostName (the /hosts key), in ~20 log lines and in exception messages that reach
DriverHealth.LastError, and the client's request-URI messages (timeouts, stream
end, not-a-stream, framing) embedded it too - while the Test-Connect probe
deliberately redacted it. New MTConnectUriRedaction (user-info, and for the
configured string its query, removed; a clean URI is unchanged so host keys do
not move); the driver computes its display form once per options document.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
I-1: a PUT/GET-disabled CPU (every read denied) stayed Faulted with
LastSuccessfulRead frozen, so the comms-loss watchdog reconnected and
re-subscribed it every 90 s forever. Both access-denied arms now also record
an answer: Faulted is kept (worst state wins), the stamp advances, and the
sticky write fault still clears only on a real success. Pinned end to end in
Runtime.Tests: the real S7Driver under the real DriverInstanceActor watchdog
is not reconnected across 2x the threshold; the control (every read timing
out) is.
Minors:
- m-1: injectable pass clock (UtcNow); "LastSuccessfulRead advanced" is now
asserted strictly, including the write-refusal case.
- m-2: the editor uses S7TagShapeRules.IsByteAnchored instead of its own list.
- m-3: a release that races an init publish also cancels the probe loop the
publish started after the release took its snapshot.
- m-4: a retained poll failing during a lifecycle operation writes no health.
- m-5: a multi-var chunk that fell back memoises its members refused alone.
- m-7: access-denied needs S7.Net's exact wording and no transport fault in
the chain.
- m-8: sync teardown clears the write-side Degraded owner; a sync Dispose
racing a running shutdown is tested.
Refs #742
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The connectivity probe parsed the full /probe device model every 5 s and discarded
it. It now calls a new IMTConnectAgentClient.PingAsync: one GET /probe read with
ResponseHeadersRead and a status check; the body is dropped unread. Same failure
surface as ProbeAsync (the default implementation delegates to it). The probe-tick
tests now count pings and pin that no tick issues a full /probe; the shutdown
test's absence assertion gained a positive control.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Key-value and where-pair plans had no row limit, so a selector pointed at a history
or events table materialised every matching row on every poll. ISqlDialect gains
RowLimitPrefix/RowLimitSuffix(n) (SQL Server TOP (n), Postgres LIMIT n, Oracle FETCH
FIRST n ROWS ONLY); plans carry a RowCap (distinct keys, or 1) and ask for RowCap+1
rows; the reader also stops client-side at RowCap+1. Getting the extra row proves the
selector ambiguous: the group publishes BadConfigurationError with a rate-limited
warning. The reader reports those answered refusals so the driver counts them as
answer evidence, and any refusal-only poll now restores Healthy after a comms fault.
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Writes were one POST per slot although the Agent's PUT surface accepts several
id=value pairs. The slots surviving the local refusals now go out as one
WriteObservationsAsync(device, pairs) per device (new IMTConnectAgentClient method;
the default loops the single-pair call). A batch closes at another device or a
repeated DataItem so same-item writes keep request order; a refused multi-slot
request is retried per slot so one refused DataItem costs only its own slot.
Health contract: any Agent answer (accepted or refused) advances
LastSuccessfulRead with the state following the read / stream flags.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Subscription throughput was hard-capped at SampleCount / SampleIntervalMs with no
lag detection: a machine producing more fell behind the ring buffer forever,
re-baselining every buffer-length with every value backlog-stale.
- The parser carries the header's optional lastSequence; each chunk measures
lag = lastSequence + 1 - nextSequence and whether it was saturated (full count).
- Three consecutive saturated chunks with a positive (or unknown) lag reopen the
stream at the cursor reached with count doubled, up to 16x SampleCount (max
65536), through a new IMTConnectAgentClient.SampleAsync(from, count, ct) overload
(default implementation ignores count). A raised stream that delivers nothing
steps back to the last good count; at the ceiling one Warning per pump.
- MTConnectDriver implements IDriverDiagnostics: sampleCount, sampleLag,
sampleSaturatedChunks, sampleCountRaises on the /hosts bag. The cap is
documented on SampleCount.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
WriteAsync published one awaited QoS 1 round-trip per request, so 50 writes to a
stalled broker cost 50 publish deadlines. Different topics now publish concurrently
(at most 8 in flight); same-topic writes stay in request order. A caller cancel still
fills every slot. A successful publish advances LastSuccessfulRead. MiniBroker now
frames packets properly and records/PUBACKs PUBLISHes.
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- T-4: every ReadAsync batch fetched and parsed the whole device /current and
built a fresh index even while the /sample pump held a live one. A read is now
answered from the pump's index (read only - the pump stays its single writer)
while the current-generation pump is running, has delivered since it started and
is not degraded; /current stays the fallback, and its per-batch index is built
over a tag map frozen once per binding instead of per call. A served read clears
the read-path degradation a failed /current left.
- T-2: the /current fallback checks the answer's instanceId; a change (an Agent
restart seen by a browse-only driver) drops the cached /probe model and raises
rediscovery once per distinct id, only while the read's session is still installed.
- The pump also acts on no verdict (Degrade / latch) once its session generation
is gone, so an abandoned pump cannot write health over the new session (T-1 follow-through).
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
A pump the lifecycle retired or abandoned that finished after the restart
installed the new session published its stale cursor onto it (PublishAgentCursor
only checked for a null or identical session), so the next pump read its first
chunk as an Agent restart: index cleared, a Bad flicker on every subscribed tag,
a needless re-baseline. The pump now carries the session generation it was
started against; the cursor publish is checked under _subscriptionLock (where the
session install and generation bump are atomic) and a superseded pump also stops
folding late chunks / re-baselines into the new session's index.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The ingestor refuses array metrics and unmapped types, yet DescribeMetric still stated
their binding tuple, so a browse-commit produced a tag stuck at
BadWaitingForInitialData forever. Those leaves now state no address (the commit
refuses them) and say why in the descriptor line.
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The authored window was persisted and consumed by nothing. It now (clamped 1-120 s)
extends the Sparkplug OpenBudget, bounds a passive UntilStable-style settle on open
(return after a 1 s quiet spell in NBIRTH/DBIRTH traffic), and bounds how long an
operator rebirth request waits for the births it asked for before returning, so the
re-rendered tree carries them. A caller deadline during the wait still reports the
publish.
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- The background probe reported the host Stopped whenever a read pass held
_gate past the probe deadline: contention is the connection being used,
not lost. A contended tick now gives no verdict (counted for tests); a
real probe failure still reports Stopped. A tick that races a teardown no
longer reopens a connection on an uninitialised driver.
- S7DriverProbe deserialised S7DriverOptions directly, whose Timeout is a
TimeSpan no config names, so "timeoutMs" was ignored and every handshake
waited the whole Test Connect budget. The handshake and test read are now
bounded by timeoutMs, capped by the caller's timeout.
Refs #742
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
- AttributesAsync was a stub returning nothing; it now reads NodeClass /
DisplayName / DataType / ValueRank / UserAccessLevel in one request and maps them
through UpstreamAttributeMapping, the same mapping discovery registers with (moved
into the shared .Session leaf; the driver forwards to it). Objects have no value
attributes.
- Certificate auth was refused "in v1". The browse now builds its identity through
OpcUaUserIdentityFactory (shared .Session leaf, the driver forwards to it),
including the secret: resolution of Password / UserCertificatePassword through the
host's ISecretResolver (optional ctor param, DI-injected). Fail-closed: a secret:
ref with no store throws instead of being sent as the literal (the browse used to
send a secret: Password verbatim). The identity is built before any network I/O.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
A span's block is the union of its members' ranges, so one member that runs
past the DB's real end (or over-declares a STRING) made the CPU refuse the
whole block, and every neighbour within it was Bad on every poll. The
planner's stated invariant (a grouped read cannot fault where per-tag reads
would succeed) was false; the multi-var path already fell back, spans did not.
- A refused multi-member span hands its members back; they are read one at a
time for that pass, bypassing the multi-var pass.
- _noGroup remembers the members refused on their own (or every member when
none was) and keeps them out of later plans, so the rest group again. It is
cleared by every initialize publish. Non-refusal span failures are
unchanged (every member stamped).
- The once-per-tag refusal Warning memo (R07:H4) is now also cleared by the
publish, and shutdown resets the write-side Degraded owner, as documented.
- Live gate: MW10 + MD254 (2 bytes past the 256-byte M area) merged into one
refused block; MW10 keeps reading 4097 on every pass.
Refs #742
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The knob was authored by the AdminUI and parsed, but consumed by nothing. It now
caps the probe deadline (shorter of the caller's budget and the config; never
extends it; 1 s floor), the same rule as the Galaxy probe.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Health contract: a write response (per-node refusals included) is the server
answering, so it restores Healthy and advances LastSuccessfulRead; a write with no
reply (transport failure / timeout) writes Degraded with the stamp preserved; a
caller-cancelled write touches nothing. A good keep-alive with no outage or dead
subscription latched now clears a Degraded left by a failed read/write instead of
only moving the stamp (a subscription-only driver otherwise stayed Degraded until
its next re-initialise). All writes are generation-stamped.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
No read or write loop checked _plcDead between items: on a frozen-but-
established peer the first op blew its deadline and marked the handle dead,
and every remaining span / tag / write still went to the wire and could wait
its own deadline under _gate (a 20-span pass ~100 s), with Degraded and the
comms-loss watchdog's clock deferred until the pass ended.
- The span loop, the multi-var chunk loop, the per-tag loop and the write
loop stamp BadCommunicationError without I/O once _plcDead is set.
- A multi-var chunk that dies on a transport loss (socket / IO / timeout /
ConnectionError) marks the handle dead instead of falling back to one
doomed per-tag read per member. The framing class is deliberately excluded
there: snap7/S7.Net report a multi-var with one refused item as
ReadData(WrongNumberReceivedBytes) over a healthy stream.
- Live gate via an in-test freezable TCP proxy (counts request frames, not
seconds). Mutation shows S7.Net 0.20 already self-closes on a cancelled op,
so the gate pins the end-to-end property; the driver guard itself is proven
by the unit tests (an op that ignores its token keeps the socket open).
Refs #742
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
One SemaphoreSlim serialised reads, writes, subscribes, acks and the whole recursive
discovery browse, so a stalled write (up to the operation timeout) blocked every
read and subscription create, and a discovery browse would block the data plane for
its full duration. Reads/acks keep the read gate; writes and DiscoverAsync are
ungated (the SDK session is safe for concurrent service calls, the host bounds write
concurrency); subscription management takes its own gate. Dispose drains both
(DisposeAsync's drain no longer over-releases on a timed-out wait).
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
A failed SUBSCRIBE (refusal or no broker verdict) used to drop its topic from every
retry set while SubscribeAsync returned a live handle, leaving tags dark behind a
Healthy connection until a redeploy. Failed topics now join a pending set retried by
the next SubscribeAsync and by every reconnect (not counted toward the reconnect's
total-failure verdict, so an ACL-denied topic cannot reconnect-loop). A no-verdict
subscribe degrades its refs then throws so the actor's reconcile re-subscribes. A
subscribe with no transport attached is recorded pending, never established (M-11).
MiniBroker now answers SUBSCRIBE.
Refs #744
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
AcknowledgeAsync swallowed every non-cancellation transport failure (and returned
normally when the session was lost while the ack queued), so the host logged the
Part 9 ack as applied while upstream still showed it unacknowledged. Both now throw
(ServiceResultException with the transport status / BadNotConnected). A reply,
including per-ack refusals, still returns normally and counts as the server answering.
Refs #743
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
A deterministic PLC refusal (object does not exist, address out of range,
type inconsistent...) degraded driver health, and an all-refused driver (a
wrong DB number) sat Degraded with LastSuccessfulRead frozen, so the 90 s
comms-loss watchdog reconnected it forever against a reachable PLC.
- IsRefusal / IsAccessDenied classify over the whole exception chain. Live
probe on the snap7 fixture: S7.Net's READ paths raise the item error as a
BARE Exception ("Received error from PLC: ..."), its write paths as
PlcException(WriteData). The old PlcException-only match missed the read
shape entirely, so a real read access-denied was BadCommunicationError +
Degraded instead of the Faulted PUT/GET alert; the Test Connect probe had
the same blind spot.
- A refusal (read or write) is BadDeviceFailure with a once-per-tag Warning,
no health write and no reopen; it counts as answered, so it advances
LastSuccessfulRead and restores Healthy after a comms fault. An undecodable
buffer (S7ValueRangeException) also counts as answered.
- A write batch the CPU answered without a failure advances
LastSuccessfulRead and retires a Degraded the write side itself published.
- Tests that used ReadData as their generic "device fault" now use SendData
so their Degraded leg is still exercised.
Refs #742
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
The init-time tag guards THREW: one mis-authored tag (DB1.DBZ0, DB1.DBW0 typed
Float64, T5 typed Int32, a writable Bool array) faulted the whole driver and the
actor retried the same init every 10 s forever, darkening every tag on the PLC.
Neither the editor nor the deploy gate checked any of it.
- S7AddressParser moves to S7.Contracts (same namespace) and a new
S7TagShapeRules there is the single authority on servable shapes.
- S7TagDefinitionFactory.FromTagConfig applies it, so the AdminUI editor, the
deploy gate (Inspect, with the specific reason) and the driver init refuse
the same tags. Init now skips a refused tag (ISkippedTagReporter) instead of
throwing; the throwing RejectUnsupportedTagConfigs is gone.
- New rules on top of the old guards: a narrow type at the wrong access width
(every read failed) and a Bool array not starting at bit 0 (U4: DB1.DBX0.3
silently read bit 0 of bytes 0..n).
- AdminUI test fixtures that used addresses the driver could never serve
(DB1.DBS0, a browse "SomeAddress", Int32/Float32 on DBW) now use real ones.
Refs #742
Claude-Session: https://claude.ai/code/session_01Ub9hGmYLQvn9XETzw7CMsT
Live gate A1 (docker-dev, opc-plc restart / SIGINT / kill): the fast path never fired —
a real process stop delivers no session status; the keep-alives report
BadSecureChannelClosed (0x80860000) once, then BadConnectionClosed (0x80AE0000)
interleaved with BadNoCommunication for the whole outage.
Reproduced in-process with TcpStallProxy.HardClose (reset every connection, refuse new
ones; server and session alive behind it): the identical sequence, and the SDK 1.5.378
session never opens a new connection on its own (48 s with the proxy accepting again,
zero attempts) — a closed channel cannot recover in place.
The driver now treats BadSecureChannelClosed/BadConnectionClosed as definitive on the
SECOND closed-channel keep-alive failure in one outage (a good keep-alive resets the
count), raising ConnectionLost ~1 keep-alive after the drop. BadNoCommunication /
BadRequestTimeout / BadNotConnected (stall/pause) stay with the watchdog; the stall test
still sees no ConnectionLost. New wire test (hard close + reopen ⇒ raised once, SDK never
reconnects, reinit brings the identical value back Good) and unit tests (sustained raises
once; a single blip that recovers does not) — all red on the previous code. Docs updated.
1. A per-ref delivery counter (bumped under the delivery lock by every publish path) is
captured before the trailing read is issued; a ref delivered to since is skipped.
Without source timestamps the read's sample could not be ordered against a notification
that landed mid-read, and publishing it regressed the node. Test injects the
notification inside the mocked read, with and without a read source time (red before).
2. Delivery-lock comment corrected (it serialises across subscriptions, which the SDK
dispatches on separate workers); PublishCommsLoss now publishes under the lock too.
3. CancelSessionWork cancels without disposing, so a keep-alive thread reading the old
token cannot hit ObjectDisposedException.
1. RewriteEndpointUrl keeps the advertised port when the configured URL names none
(Uri.Port -1 would have sent the session to 4840); port-less + IPv6 tests.
2. The trailing re-read repairs any (value, status) difference that is not older than
the last delivered value — Good->Bad/Uncertain included; a read without a source
timestamp counts as newer; an equal source time with different content is published
(repairs the replay guard's by-design equal-timestamp drop).
3. One driver-wide delivery lock makes decide+publish atomic across live notifications,
the recovery read and the trailing read, so an older value can never land after a
newer one (deterministic race test via a seam).
4. Known gap documented: a ConnectionLost raised while the host is not Connected is
swallowed and that session falls back to the watchdog.
5. Only a SERVER-reported subscription loss raises ConnectionLost; the publishing-stall
heuristic stays with the #527 watchdog.
6. Post-recovery timers run on Task.Delay(…, Clock, ct) with a session-work CTS
cancelled on session release and dispose.
7. Replay summary at Information only when something was dropped (Debug otherwise);
a re-arm reports the superseded window's drops and counts afresh.
8. DriverInstanceActor: AttachHostStatusSource doc comment restored; Reconnecting
swallows a stray DisconnectObserved (test).
9. Stall wire test: 120 s session timeout, 3000-cycle subscription lifetime, one further
bad keep-alive tick, 60 s headroom on the post-recovery waits (5/5 green).
- Endpoint rewrite: the driver now connects to the configured host AND port (scheme
match, advertised path kept) instead of the server-advertised URL — fixes NAT/port-map
upstreams and lets an in-test proxy carry the session (SDK SelectEndpoint rule).
- TcpStallProxy + OpcUaClientDriverStallWireTests: the in-place blip (#670) is now
wire-tested on loopback — one Bad, frozen read stamp, same session, identical value
republished Good, a post-recovery change arrives, no source-time regression after
the SDK's late backlog flush.
- Defect found by that test, fixed: a change published right after a stall rode a
Publish response the SDK timed out and never republished, leaving the node Good on
the pre-change value indefinitely. New trailing reconcile read (Timeout +
KeepAliveInterval after recovery) publishes any Good value newer than the last
accepted one, through the replay guard's mark.
- IDriverConnectionLossSource (Core.Abstractions): opt-in definitive-loss signal.
DriverInstanceActor attaches in PreStart / detaches in PostStop and turns a raise
into a disconnect only while Connected. OpcUaClient raises once per session on
keep-alive BadSessionIdInvalid/BadSessionClosed/BadSessionNotActivated/
BadServerHalted (measured: the SDK never re-opens a halted channel) or a dropped
subscription; ambiguous codes stay with the #527 watchdog.
- Replay guard: one Information summary per recovery when the window closes
("dropped N late notifications across M refs"), with a timer so quiet
subscriptions log it too.
- ReconnectPeriod deleted (options, AdminUI form, docs, tests); stored configs that
carry reconnectPeriod still parse in the factory, probe and browser.
- Docs: OpcUaClient.md, DriverLifecycle.md, driver-specs.md, CLAUDE.md.
The F2 live gate failed (10 value regressions over 4 blips): IsLateReplay
disarmed a ref on its first in-order notification, which arrives ~1 s after
recovery, while the SDK's late replay lands 4-6 s later. Equal source times
also passed, so a replay equal to the recovery read got through.
While the window is open each ref now carries a high-water mark (seeded at
arming, raised by the recovery read, advanced by every accepted notification);
a notification at or below it is dropped (equal included — OPC UA moves the
source timestamp on every value/status change). Null source time passes
without moving the mark. Window expiry still disarms entirely, so an upstream
clock step-back is never frozen outside the window. Arm/expiry serialised by a
lock so an expiry racing a re-arm cannot wipe fresh seeds.
The global 'never older than the last Good source time' rule froze a node
whenever an upstream clock stepped backwards (NTP correction, PLC clock
reset) until the clock caught up. The check is now a replay guard armed
only by a same-session recovery (keep-alive bad->good): per ref it holds
the last Good upstream source time (raised by the recovery read), a ref
disarms on its first in-order notification, and the whole guard expires
after max(30 s, 2x the longest subscription keep-alive window). Outside
it every notification is accepted. The steady-state tracker records the
LAST accepted time, not a high-water mark.
F1: the driver's outage BadNoCommunication carried the last Good sample's
source timestamp. The recorder historizes SourceTimestampUtc ??
ServerTimestampUtc, so the Bad landed on the Good row's (TagId,
SourceTimeUtc) key and the insert-if-absent historian dropped it - no
outage marker until the host's watchdog marker ~95 s in, none at all for
a short blip. Now stamped with the detection time (source + server), as
the host's comms-loss fan does.
F2: after a same-session recovery the SDK delivered a missed notification
seconds late with an older source time and the node regressed. A
notification older than the newest Good upstream source time for its ref
(notification or recovery read) is dropped and counted
(stale_notification_dropped); equal timestamps pass; the outage Bad never
moves the comparison; a re-initialise resets it.
M1: a subscription that dies while the session lives (outage longer than
the subscription lifetime but shorter than the session timeout) used to
end Healthy with fresh keep-alive stamps and frozen values. PublishError
BadNoSubscription / BadSubscriptionIdInvalid naming a tracked
subscription, or a tracked subscription whose SDK PublishingStopped
persists one more keep-alive window + 5 s across good keep-alives, now
declares it lost: Degraded, outage published once, and nothing advances
LastSuccessfulRead until the session is released, so the #527 watchdog
re-initialises and re-subscribes. Both detectors wire-tested.
M2: in-place recovery is bound to the session whose keep-alive recovered;
one overtaken by a re-initialise no longer re-arms the new session's
outage latch or degrades its health (ReadCoreAsync expectedSession).
L1 CAS health updates; L2 re-check before each recovery publish; L3 skip
CloseSession on a lost session; L4 SDK log forwarding rate-limited
(20/min); L5 per-driver SDK LoggerFactory disposed with the driver.
Wire test asserts the read stamp stays frozen while the server is down.
A bad keep-alive now degrades health (LastSuccessfulRead preserved) and,
once per outage, publishes BadNoCommunication (value retained) for every
subscribed ref plus a quality-only annotation on subscribed alarm
conditions. Good keep-alives and data/event notifications advance
LastSuccessfulRead, so an idle-but-alive subscription never looks dead
to the #527 watchdog. A keep-alive that recovers on the same session
re-reads the Bad refs and republishes them Good; a session that does not
come back is recovered by the watchdog's ShutdownAsync + InitializeAsync
+ re-subscribe.
Deletes the SessionReconnectHandler wiring (it retried forever in SDK
1.5.378, cloned subscriptions the driver never re-tracked and kept health
Healthy), OnReconnectComplete, TryRearmReconnect and the reconnect_*
counters (new keepalive_recovered). Teardown of a lost session skips
wire deletes and bounds CloseSession. SDK warnings reach the driver log.
Wire test reworked: restart => Bad, then watchdog-style reinit => Good.
- Historian.md: Empty samples always Bad (writer/updater/read clamp),
long-name refusal, 5 s bootstrap failure cache, AtTime after now, info-bit
masking, HistoryUpdate does NOT type-check (M-1 decision), DeleteRaw
needs both bounds, reverse Raw paging, the historian/historization metric
prefixes, and provisioning numeric-only lines qualified as Gateway.
- CLAUDE.md numeric-only lines qualified as Gateway backend.
- Stale 'sole historian' lines in DriverLifecycle, ServiceHosting and the
retired Wonderware stub; Client.CLI exit-code contract.
- ServerHistorianOptions: Endpoint/Backend XML docs untangled, Validate()
indented; Program.cs comments are backend-neutral; T10 week-test day count.
Claude-Session: https://claude.ai/code/session_01Ef2ozxS7QPanGT8uoATbTe
Since #516 a DriverConfig change is stop + respawn, so DriverSpawnPlan.ToApplyDelta
was always empty and the in-place path had no producer. #663 and #664 were both
latent bugs in it. This deletes it:
- host: ApplyChildDelta, _pendingDelta, HandleApplyResult and HandleDeltaApplied
- plan: DriverSpawnPlan.ToApplyDelta
- actor: the ApplyDelta, ApplyResult and DeltaApplied messages, the Connected
reinit handler and the mid-connect AdoptConfigDuringInit
- tests: the actor tests that drove the seam
IDriver.ReinitializeAsync and each driver's implementation and unit tests stay.
Comments and docs that cited AdoptConfigDuringInit as the reason a driver must
release prior state at init now cite the IDriver contract and overlapping
retries, which still apply.
Claude-Session: https://claude.ai/code/session_01Ef2ozxS7QPanGT8uoATbTe
Joseph Doherty