Commit Graph

2 Commits

Author SHA1 Message Date
Joseph Doherty
d11dd0520b Galaxy IPC unblock — live dev-box E2E path
Three root-cause fixes to get an elevated dev-box shell past session open
through to real MXAccess reads:

1. PipeAcl — drop BUILTIN\Administrators deny ACE. UAC's filtered token
   carries the Admins SID as deny-only, so the deny fired even from
   non-elevated admin-account shells. The per-connection SID check in
   PipeServer.VerifyCaller remains the real authorization boundary.

2. PipeServer — swap the Hello-read / VerifyCaller order. ImpersonateNamedPipeClient
   returns ERROR_CANNOT_IMPERSONATE until at least one frame has been read
   from the pipe; reading Hello first satisfies that rule. Previously the
   ACL deny-first path masked this race — removing the deny ACE exposed it.

3. GalaxyIpcClient — add a background reader + single pending-response
   slot. A RuntimeStatusChange event between OpenSessionRequest and
   OpenSessionResponse used to satisfy the caller's single ReadFrameAsync
   and fail CallAsync with "Expected OpenSessionResponse, got
   RuntimeStatusChange". The reader now routes response kinds (and
   ErrorResponse) to the pending TCS and everything else to a handler the
   driver registers in InitializeAsync. The Proxy was already set up to
   raise managed events from RaiseDataChange / RaiseAlarmEvent /
   OnHostConnectivityUpdate — those helpers had no caller until now.

4. RedundancyPublisherHostedService — swallow BadServerHalted while
   polling host.Server.CurrentInstance. StandardServer throws that code
   during startup rather than returning null, so the first poll attempt
   crashed the BackgroundService (and the host) before OnServerStarted
   ran. This race was latent behind the Galaxy init failure above.

Updates docs that described the Admins deny ACE + mandatory non-elevated
shells, and drops the admin-skip guards from every Galaxy integration +
E2E fixture that had them (IpcHandshakeIntegrationTests, EndToEndIpcTests,
ParityFixture, LiveStackFixture, HostSubprocessParityTests).

Adds GalaxyIpcClientRoutingTests covering the router's
request/response match, ErrorResponse, event-between-call, idle event,
and peer-close paths.

Verified live on the dev box against the p7-smoke cluster (gen 6):
driver registered=1 failedInit=0, Phase 7 bridge subscribed, OPC UA
server up on 4840, MXAccess read round-trip returns real data with
Status=0x00000000.

Task #112 — partial: Galaxy live stack is functional end-to-end. The
supplied test-galaxy.ps1 script still fails because the UNS walker
encodes TagConfig JSON as the tag's NodeId instead of the seeded TagId
(pre-existing; separate issue from this commit).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-24 16:30:16 -04:00
Joseph Doherty
a7126ba953 Phase 2 — port MXAccess COM client to Galaxy.Host + MxAccessGalaxyBackend (3rd IGalaxyBackend) + live MXAccess smoke + Phase 2 exit-gate doc + adversarial review. The full Galaxy data-plane now flows through the v2 IPC topology end-to-end against live ArchestrA.MxAccess.dll, on this dev box, with 30/30 Host tests + 9/9 Proxy tests + 963/963 solution tests passing alongside the unchanged 494 v1 IntegrationTests baseline. Backend/MxAccess/Vtq is a focused port of v1's Vtq value-timestamp-quality DTO. Backend/MxAccess/IMxProxy abstracts LMXProxyServer (port of v1's IMxProxy with the same Register/Unregister/AddItem/RemoveItem/AdviseSupervisory/UnAdviseSupervisory/Write surface + OnDataChange + OnWriteComplete events); MxProxyAdapter is the concrete COM-backed implementation that does Marshal.ReleaseComObject-loop on Unregister, must be constructed on an STA thread. Backend/MxAccess/MxAccessClient is the focused port of v1's MxAccessClient partials — Connect/Disconnect/Read/Write/Subscribe/Unsubscribe through the new Sta/StaPump (the real Win32 GetMessage pump from the previous commit), ConcurrentDictionary handle tracking, OnDataChange event marshalling to per-tag callbacks, ReadAsync implemented as the canonical subscribe → first-OnDataChange → unsubscribe one-shot pattern. Galaxy.Host csproj flipped to x86 PlatformTarget + Prefer32Bit=true with the ArchestrA.MxAccess HintPath ..\..\lib\ArchestrA.MxAccess.dll reference (lib/ already contains the production DLL). Backend/MxAccessGalaxyBackend is the third IGalaxyBackend implementation (alongside StubGalaxyBackend and DbBackedGalaxyBackend): combines GalaxyRepository (Discover) with MxAccessClient (Read/Write/Subscribe), MessagePack-deserializes inbound write values, MessagePack-serializes outbound read values into ValueBytes, decodes ArrayDimension/SecurityClassification/category_id with the same v1 mapping. Program.cs selects between stub|db|mxaccess via OTOPCUA_GALAXY_BACKEND env var (default = mxaccess); OTOPCUA_GALAXY_ZB_CONN overrides the ZB connection string; OTOPCUA_GALAXY_CLIENT_NAME sets the Wonderware client identity; the StaPump and MxAccessClient lifecycles are tied to the server.RunAsync try/finally so a clean Ctrl+C tears down the COM proxy via Marshal.ReleaseComObject before the pump's WM_QUIT. Live MXAccess smoke tests (MxAccessLiveSmokeTests, net48 x86) — skipped when ZB unreachable or aaBootstrap not running, otherwise verify (1) MxAccessClient.ConnectAsync returns a positive LMXProxyServer handle on the StaPump, (2) MxAccessGalaxyBackend.OpenSession + Discover returns at least one gobject with attributes, (3) MxAccessGalaxyBackend.ReadValues against the first discovered attribute returns a response with the correct TagReference shape (value + quality vary by what's running, so we don't assert specific values). All 3 pass on this dev box. EndToEndIpcTests + IpcHandshakeIntegrationTests moved from Galaxy.Proxy.Tests (net10) to Galaxy.Host.Tests (net48 x86) — the previous test placement silently dropped them at xUnit discovery because Host became net48 x86 and net10 process can't load it. Rewritten to use Shared's FrameReader/FrameWriter directly instead of going through Proxy's GalaxyIpcClient (functionally equivalent — same wire protocol, framing primitives + dispatcher are the production code path verbatim). 7 IPC tests now run cleanly: Hello+heartbeat round-trip, wrong-secret rejection, OpenSession session-id assignment, Discover error-response surfacing, WriteValues per-tag bad status, Subscribe id assignment, Recycle grace window. Phase 2 exit-gate doc (docs/v2/implementation/exit-gate-phase-2.md) supersedes the partial-exit doc with the as-built state — Streams A/B/C complete; D/E gated only on the legacy-Host removal + parity-test rewrite cycle that fundamentally requires multi-day debug iteration; full adversarial-review section ranking 8 findings (2 high, 3 medium, 3 low) all explicitly deferred to Stream D/E or v2.1 with rationale; Stream-D removal checklist gives the next-session entry point with two policy options for the 494 v1 tests (rewrite-to-use-Proxy vs archive-and-write-smaller-v2-parity-suite). Cannot one-shot Stream D.1 in any single session because deleting OtOpcUa.Host requires the v1 IntegrationTests cycle to be retargeted first; that's the structural blocker, not "needs more code" — and the plan itself budgets 3-4 weeks for it.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-18 00:23:24 -04:00