CentralDbTestEnvironment sets five process-wide environment variables, and
Program's AddEnvironmentVariables() reads them at an unpredictable point during
host boot. With xUnit collection parallelization on, one fixture's teardown
could clear a var mid-boot for a sibling. Since the secrets adoption (G-4)
three of those keys are ${secret:...} references that fail closed, turning a
previously benign empty value into a SecretNotFoundException that aborts the
boot — an intermittent CI failure.
Adds a HostBootCollection that serializes every fixture booting a real host
while depending on that shared state, folding in the narrower "ActorSystem"
collection so its members stay serialized with each other as before. Site-role
fixtures stay parallel: they call Configuration.Sources.Clear(), dropping the
env-var provider, so they cannot participate in the race.
CentralDbTestEnvironment now also fails fast if two instances are ever live at
once, making a regression (a fixture added outside the collection) deterministic
rather than intermittent — this is what surfaced the disposed-CTS defect fixed
in the previous commit. Fixture teardown is try/finally so a throwing host
teardown can no longer strand the vars for the rest of the run.
Cost: Host.Tests runs ~2m30s -> ~4m10s; the serialized Central-boot classes are
most of the assembly's parallelism.
Fixes: Gitea #15
REQ-HOST-4a lists "required cluster singletons running (if applicable)" as a
readiness criterion, but /health/ready only checked database + akka-cluster.
Add a third Ready-tagged check, RequiredSingletonsHealthCheck, registered in the
Central-role AddHealthChecks() chain (so it is naturally role-scoped — site nodes
never run it).
Probe: for each required central singleton, Ask its local ClusterSingletonProxy
an Identify with a short bounded per-singleton timeout (~2s, probes run
concurrently via Task.WhenAll). A non-null ActorIdentity.Subject within the
timeout means the singleton is running and reachable through the proxy; a null
subject or a timeout means unreachable → Unhealthy, naming the unreachable
singleton(s). The check never throws (catch-all → Unhealthy) and resolves
ActorSystem lazily from DI per probe (Unhealthy if Akka not yet up).
Required-always set = the five singleton proxies created unconditionally in
AkkaHostedService.RegisterCentralActors: notification-outbox, audit-log-ingest,
site-call-audit, audit-log-purge, site-audit-reconciliation. There are no
feature/config-gated central singletons today; any future gated singleton is the
"if applicable" case and must NOT be added to the required set.
Leadership-agnostic: the proxy reaches the singleton from either central node, so
a ready standby still reports ready (readiness must not require cluster
leadership — that is the Active tier's job). During a brief singleton handover the
probe may time out and the node flaps to not-ready, which is correct (a node
mid-handover is legitimately not fully ready); no retries, to keep the probe fast.
Tests (TDD): RequiredSingletonsHealthCheckTests exercises the probe against a
TestKit ActorSystem — all proxies present+reachable → Healthy; one missing →
Unhealthy naming it; ActorSystem absent → Unhealthy, no throw. HealthCheckTests
regression-guards the Ready tag + absence of the Active tag on the new check.
Joseph Doherty