- CreateDefaultHttpClient sets AllowAutoRedirect=false so a 307/308 cannot re-POST the
/auth/token credential form elsewhere.
- A forced re-exchange after a 401 carries the refused token; inside the gate it reuses
a token another caller already re-exchanged instead of binding again. The BasicOnly
switch is written under the same gate.
- SCADABRIDGE_VERBOSE=1 prints a stderr note when a fresh token is refused and the CLI
falls back to Basic.
- debug stream's AccessTokenProvider retries the exchange once and documents the
null-token degradation (the hub refuses, automatic reconnect retries).
- Tests: concurrent refusal reuses the peer refresh (2 exchanges, not 3); real-socket
redirect is not followed.
- JwtTokenService back-dates nbf by the 30 s validation skew so a peer central whose
clock lags the issuer accepts a fresh token after failover.
- RoleMappingCacheSeconds = 0 routes RoleMapper to the matched-rows loader (old cost)
instead of loading every Deployer mapping's scope rules per call.
- RoleMappingCache / SecurityOptions / Component-Security.md describe Invalidate() by
its contract plus the TTL, not by a call site.
- Tests: HS512 with the same key, issuer/audience mismatch, lagging-peer nbf, TTL-0
matched-only reads, refresh-vs-Invalidate generation race.
- ManagementAuthenticator accepts Authorization: Bearer <jwt> alongside Basic on
/management, the audit REST endpoints and the debug-stream hub. Tokens are
validated locally (HS256 only, alg=none rejected, issuer/audience, lifetime with
30 s skew); roles and site scope inside a valid token are authoritative for its
lifetime, so a bearer call costs no LDAP bind and no role re-map.
- /auth/token returns expires_in + management_bearer:true, carries username, display
name, roles and permitted sites, and answers 404 under DisableLogin.
- CLI exchanges credentials once per process (in-memory CliTokenStore), sends Bearer,
re-exchanges once on 401 then falls back to Basic; older/login-disabled servers are
Basic-only; bad credentials are not retried with Basic. Debug stream uses the token
via AccessTokenProvider. SCADABRIDGE_AUTH_MODE=basic opts out.
- StartupValidator fails a Central boot on a JwtSigningKey under 32 UTF-8 bytes;
fixed the false 'fail-fast at construction' comment. Removed the dead
RefreshToken/RecordActivity/ShouldRefresh/IsIdleTimedOut members and the inert
JwtRefreshThresholdMinutes; JwtExpiryMinutes now documented as the bearer lifetime
(validated > 0).
- RoleMapper reads through a process-wide RoleMappingCache (RoleMappingCacheSeconds,
default 30 s): one refresh loads the table plus every Deployer mapping's scope
rules, Invalidate() drops it (generation-guarded against in-flight refreshes).
- Component-Security.md / Component-CLI.md / CLI README describe the real model
(cookie claims ticket, bearer for CLI) instead of a cookie-embedded JWT.
GetAuthenticationStateAsync (display reads: current username, nav role checks) used
to force a full revalidation + DB role re-map on every call. It now reuses the
circuit's last revalidation for NonAuthoritativeRefreshInterval (10 s);
GetAuthoritativeStateAsync (mutation gates, UI-13) still re-maps every time and
fails closed. The revocation regression test now asserts through the authoritative
read, which is what every gate uses.
InstanceService.SetAlarmOverrideAsync and TemplateDeletionService move to the
no-tracking, body-free GetAllTemplatesForAnalysisAsync; TemplateFolderService
counts folder contents from GetTemplateStubsAsync.
Advisory, not a trust-policy violation: every policy entry blocks deploy, and
an error here would fail the next redeploy of already-deployed scripts.
ValidationService surfaces it as a non-blocking ScriptCompilation warning.
RoslynScriptCompiler.Compile now uses Script.GetCompilation().GetDiagnostics()
with one shared, never-loading InteractiveAssemblyLoader instead of
Script.Compile(), which emitted and loaded a non-collectible assembly (and a
fresh AssemblyLoadContext) per distinct script. The catch-all that turned any
exception into a 'Compilation exception' diagnostic is gone; TryCompile and
CheckExpressionSyntax report a fault for that call only, uncached.
Targeted transitive pins per the house pattern, no family-wide bump, no suppression:
Cryptography.Xml 10.0.10->10.0.13; Pkcs pinned 10.0.13 in ConfigurationDatabase,
ExternalSystemGateway, NotificationService, Playwright tests; SignalR.Common 10.0.13 in CLI.
Comment/XML-doc changes only:
- add summaries to ~10k test methods, fixtures, fakes and helpers
(+ <param> on theories, <returns> on async Task methods,
<inheritdoc/> on interface implementations the checker can resolve)
- strip ~2.6k project-tracking references (WP*, P*-*, row-NN, Task N,
#NN, Component-0NN) from comments, keeping the explanatory text;
reword false positives in comments only
- real docs on fakes implementing BCL/library interfaces
(IDisposable, ILogger, IOptionsMonitor, IAsyncLifetime)
- reword the last 5 src/ false positives (CWE-1236, failures-1,
year-0001, M365)
Only non-comment edits: four one-line nested classes split across lines
so members can carry docs. Build 0 errors / 0 warnings.
Comment/XML-doc changes only, no code logic:
- strip project-tracking references (WP*, P*-*, row-NN, #NN, Task N) from
comments, keeping the explanatory rationale
- add missing summaries, <param>, <returns>, <typeparam> and <inheritdoc/>
- replace misused <inheritdoc/> with real docs; where <inheritdoc/> was
combined with useful remarks, move the rationale into // comments
- drop stale replicationActor <param> on DeploymentManagerActor ctor
Remaining 5 are false positives (CWE-1236 x2, failures-1, year-0001, M365).
Build 0 errors / 0 warnings.
Rig finding on the upgrade restart: both nodes' rows were older than the stale threshold
(the old build never refreshed them), so node-a's reporter pruned node-b's row while node-b
wrote a fresh one, and the replicated delete won. The reporter now prunes nothing during its
first refresh interval after start-up, and each mirror pass re-writes its own row if it no
longer holds what the mirror last wrote.
Each node now re-writes its mirror-status row every SecretMirrorStatusRefreshInterval
(10 min) as a heartbeat; SecretMirrorStatusReporter leaves off the health report and
deletes any row not refreshed for SecretMirrorStatusStaleAfter (1 h, validated >= 2x the
refresh). The delete is conditional on the checked_at it read, so a node that came back
keeps its fresh row; a returning node re-writes its row on its first pass.
Also: DeploymentDeadlineParityTests' stub replies with a real zero-failure summary (the artifact
aliases now derive their exit code from failureCount), and register row 127 reflects the merge
and the two new replicated site tables.
A DeployedSecretsMirror failure on the STANDBY node (e.g. a master-key mismatch
that stops a deployed secret decrypting there) was only logged on that node:
site health reports are sent by the active node only, so central never saw it.
- New 12th LocalDb-replicated table secret_mirror_status (node_name PK,
failures_json of {SecretName, Message}, checked_at; no FKs). Each node's
mirror writes only its own row, and only when its failure set changes plus
once per start-up (every write replicates via CDC). Best-effort: a failed
write is a Warning and is retried next pass. Node key = INodeIdentityProvider
NodeName, falling back to the machine name.
- SecretMirrorStatusReporter (SiteRuntime, every node, 30 s) reads every row
into ISiteHealthCollector.SetSecretMirrorStatuses -> additive nullable
SiteHealthReport.SecretMirrorStatuses (null = no data).
- Wire: SiteHealthReportDto.secret_mirror_statuses = 33 (presence wrapper so
null != empty); vendored CentralControl.cs regenerated; mapper both ways.
- Central UI Health card: 'Secret mirror failing' badge + compact per-node
node / secret / message list.
- Tests: mirror status write-on-change/first-pass/failure isolation, storage
round-trip, reporter/collector, DTO mapper round-trip incl. null/empty,
bUnit, replicated-table set (twelve) + FK guard wording.
Secret-store outage (final review M5): SiteSecretsCollector returns a store
read failure as SiteSecretsCollection.StoreError instead of throwing. A deploy
now fails only the sites that reference a secret (nothing sent, not stamped,
DeploySiteSecrets row with the error), and ComputeCurrentArtifactHashesAsync
omits only those sites (they read Out of date) instead of the exception
aborting the whole sweep and the fleet-wide pending-updates check.
Removal-only deploys: the site reply gains SecretsRemoved (proto field 7,
names only) - the deployed_secrets rows the deploy soft-removed. Central
carries it on SiteArtifactResult and writes the DeploySiteSecrets row with a
removed list whenever it is non-empty, so dropping a site's last secret:
reference is audited centrally rather than only in site logs.
site deploy-artifacts and deploy artifacts rendered the per-site summary and
always exited 0, so a script could not tell a failed site (unreachable,
missing or unreadable secret, partial apply) from a clean deploy. Both now
route through the failureCount rule deploy site already used.
Joseph Doherty