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.
RedundancyStateActor derived the driver Primary from ClusterState.RoleLeader("driver").
Akka offers two different notions of a "first" member and they are not the same
one: role leader is the lowest-ADDRESSED Up member (host, then port), while
ClusterSingletonManager places singletons on the OLDEST — lowest up-number.
They agree on a freshly-formed cluster, which is why every existing test passed.
They diverge after any restart: the restarted node re-joins as the youngest while
keeping its address, so if it holds the lower address it becomes role leader while
the singletons stay put. The snapshot would then name a Primary that is not
hosting the work, and every Primary-gated surface follows it — inbound device
writes, native-alarm acks, the fleet-wide alerts emit, and the alarm-history
drain would all enable on the wrong node while the node actually running the
singletons stayed gated off.
BuildSnapshot now selects the oldest Up member carrying the driver role, matching
singleton placement. Leaving members are excluded: a node handing its singletons
over must not be named Primary.
NodeRedundancyState.IsRoleLeaderForDriver is renamed IsDriverPrimary, and
NodeHealthInputs.IsDriverRoleLeader likewise. Keeping the old names would have
left the wire contract asserting a derivation the code no longer uses — the same
drift that made this defect invisible.
Proven by a real two-node cluster rather than a mock. RedundancyPrimaryElectionTests
binds the first-joining node to the HIGHER port, so oldest and lowest-address name
different nodes, and includes a fixture assertion that the divergence actually
occurred — without it the real assertion could pass for the wrong reason. Positive
control: restoring the RoleLeader derivation turns exactly the two election tests
red while the fixture check stays green.
Runtime.Tests 440 passed, ControlPlane.Tests 82, Cluster.Tests 36.
Claude-Session: https://claude.ai/code/session_01GASWkNEi68FSCtvr6rLoEW
Layer 3 of #477: a scripted alarm's condition Quality now reflects the WORST
quality across its input tags, mirroring the native OT semantic (#477 L2).
Plumbing (quality was silently discarded twice on the live path):
- VirtualTagActor.DependencyValueChanged gains Quality (defaulted Good); the
DependencyMuxActor forwards the published AttributeValuePublished.Quality it
already carried; ScriptedAlarmHostActor.OnDependencyChanged pushes the real
quality into the engine (was hardcoded 0u/Good).
Engine (Core.ScriptedAlarms):
- ScriptedAlarmEngine computes worst-of-input quality each eval (skipping
not-yet-published inputs, which are a readiness concern, not a quality signal)
and carries it on ScriptedAlarmEvent.WorstInputStatusCode.
- A real transition carries the current worst quality so ToSnapshot's full
snapshot doesn't clobber quality back to Good (e.g. transition while Uncertain).
- A Bad input freezes the condition (no transition), like a comms-lost native
driver; a quality-bucket change with no transition emits the new
EmissionKind.QualityChanged, routed to the existing #477-L2
AlarmQualityUpdate -> WriteAlarmQuality node path (quality only, no /alerts
row, no historian write). ScriptedAlarmSource skips QualityChanged so it never
fabricates a phantom IAlarmSource event.
Host: ToSnapshot maps WorstInputStatusCode -> OpcUaQuality; OnEngineEmission
routes QualityChanged out of band.
Tests (TDD, RED-first): engine worst-carry + Bad/restore QualityChanged +
unchanged-bucket-no-emit; source swallows QualityChanged; mux forwards quality;
host Bad-dep -> AlarmQualityUpdate(no alerts) + transition snapshot carries worst.
Docs: AlarmTracking.md Layer-3 section + design doc.
Closes#478
H1 (HIGH): write-routing key now (AddressSpaceRealm, bareId), not bare-only.
A raw s=<RawPath> and a UNS s=<Area/Line/Equip/Eff> can collide as bare
strings; the bare-only key let a colliding raw+UNS pair route to the WRONG
driver ref (last-writer-wins). The realm the node manager resolves (RealmOf)
is now threaded through IOpcUaNodeWriteGateway.WriteAsync -> RouteNodeWrite ->
_driverRefByNodeId keyed by (realm, bareId). New regression test:
Colliding_raw_and_uns_bare_ids_route_to_their_own_driver_by_realm.
M1 (MEDIUM): discovered-node injection made coherently DORMANT. HandleDiscoveredNodes
hard-short-circuits (single enforcement point; _discoveredByDriver never
populates so the re-inject tail is inert too), with a clear log pointing at the
/raw browse-commit flow. New pin: Discovered_nodes_are_ignored_dormant_in_v3;
the 16+2 v2 injection scenarios re-pointed to an accurate skip reason
(DiscoveryInjectionDormantV3).
M2 (MEDIUM): realm-qualified dual-node self-correction tests —
Failed_uns_write_reverts_uns_node_and_leaves_raw_node_untouched +
Raw_realm_revert_reverts_raw_node_only (the second fails if the realm is dropped).
L1: removed the = AddressSpaceRealm.Uns defaults from the consequential
node-manager mutation methods (WriteValue/WriteAlarmCondition/MaterialiseAlarmCondition/
EnsureFolder/EnsureVariable/UpdateFolderDisplayName/UpdateTagAttributes/
RaiseNodesAddedModelChange/Remove*/RevertOptimisticWriteIfNeeded) + the
AttributeValueUpdate/AlarmStateUpdate records, so the compiler forces explicit
realm; read-only accessors + internal builders retain their defaults.
L3: fixed the stale VirtualTagHostActor class comment (V3NodeIds.Uns, not the
retired EquipmentNodeIds.Variable).
Also: DeploymentArtifactRawUnsParityTests — Raw/UNS node-set byte-parity
round-trip between AddressSpaceComposer.Compose and DeploymentArtifact.ParseComposition.
Claude-Session: https://claude.ai/code/session_01LVneM3eh1UtJxEisFXgmox
Validate AddComment up-front (IsNullOrWhiteSpace guard + Warning log) so
a blank-comment command is cleanly rejected before reaching the engine
rather than faulting inside ApplyAddComment and being silently swallowed
by the outer catch. Mirrors the existing TimedShelve missing-UnshelveAtUtc
pattern.
Also fix two stale inline comments: the "async void crash" note on
TimedShelve now correctly says "fault escaping async Task → supervision
restart", and the ownership-filter now documents the benign race with a
concurrent LoadAsync clearing the loaded set.
Tests: AlarmCommand_add_comment_empty_text_is_rejected_not_driven (Theory
— empty string + whitespace) and AlarmCommand_add_comment_nonempty_drives_engine
(positive path, asserts CommentAdded transition on alerts topic).
Subscribe the host to the cluster alarm-commands DPS topic in PreStart and
drive the matching ScriptedAlarmEngine op per inbound AlarmCommand. An
ownership filter (engine.LoadedAlarmIds) ignores commands for alarms this
node does not own; TimedShelve without UnshelveAtUtc and unknown operations
are logged + rejected (never thrown); op failures are caught + logged so a
faulting op can't fault the actor. Re-projection is left to the engine's
existing OnEvent -> OnEngineEmission path.
Handler is a Task-returning ReceiveAsync (the project's AK2003 analyzer
forbids an async-void Receive delegate), giving ordered awaited async on the
actor thread. Adds 3 TestKit tests: ack drives the engine with mapped args,
unowned command ignored, missing-UnshelveAtUtc TimedShelve rejected not
thrown.
Joseph Doherty