feat(opcua): emit Bad blip + AuditWriteUpdateEvent + sync fail-fast on failed device write

This commit is contained in:
Joseph Doherty
2026-06-19 02:14:58 -04:00
parent e047af0553
commit 36eb14e88d
2 changed files with 227 additions and 2 deletions
@@ -1064,6 +1064,8 @@ public sealed class OtOpcUaNodeManager : CustomNodeManager2
/// </list>
/// Silent value-wise — this node manager carries no logger; the gateway logs the underlying write failure
/// and the SDK trace captures any audit-report failure.
/// <para><c>internal</c> (not private) so the behavioural test drives the blip-then-settle continuation
/// against a booted node manager — see <c>NodeManagerWriteRevertTests</c>.</para>
/// </summary>
/// <param name="nodeId">The string id of the written variable node.</param>
/// <param name="outcome">The device-write outcome routed back by the gateway.</param>
@@ -1072,7 +1074,7 @@ public sealed class OtOpcUaNodeManager : CustomNodeManager2
/// <param name="priorStatus">The node's real status captured before the optimistic write.</param>
/// <param name="clientUserId">The writing principal's user-id string (the identity's DisplayName), threaded
/// from <see cref="OnEquipmentTagWrite"/> to populate the audit event's ClientUserId; null when unknown.</param>
private void RevertOptimisticWriteIfNeeded(
internal void RevertOptimisticWriteIfNeeded(
string nodeId, NodeWriteOutcome outcome, object? optimisticValue, object? priorValue, StatusCode priorStatus,
string? clientUserId)
{
@@ -1132,7 +1134,10 @@ public sealed class OtOpcUaNodeManager : CustomNodeManager2
/// <param name="priorValue">The node's real pre-write value (the audit OldValue).</param>
/// <param name="clientUserId">The writing principal's user-id string; null when unknown.</param>
/// <returns>A populated, unreported <see cref="AuditWriteUpdateEventState"/>.</returns>
private AuditWriteUpdateEventState BuildWriteFailureAuditEvent(
/// <remarks><c>internal</c> (not private) so <c>NodeManagerWriteRevertTests</c> can assert the populated
/// audit fields at the nearest deterministic seam (the end-to-end <c>Server.ReportEvent</c> dispatch would
/// need a subscribed event monitored-item to observe).</remarks>
internal AuditWriteUpdateEventState BuildWriteFailureAuditEvent(
BaseDataVariableState variable, NodeWriteOutcome outcome, object? optimisticValue, object? priorValue,
string? clientUserId)
{
@@ -0,0 +1,220 @@
using Microsoft.Extensions.Logging.Abstractions;
using Opc.Ua;
using Shouldly;
using Xunit;
using ZB.MOM.WW.OtOpcUa.Commons.OpcUa;
namespace ZB.MOM.WW.OtOpcUa.OpcUaServer.Tests;
/// <summary>
/// Write-outcome self-correction — the BEHAVIOURAL half of the failed-device-write surfacing
/// (the pure decision tables live in <see cref="EquipmentWriteGateTests"/>). Boots a real
/// <see cref="OtOpcUaSdkServer"/> through <see cref="OpcUaApplicationHost"/> (the same harness
/// <see cref="NodeManagerHistorizeTests"/> uses) so the continuation runs against a live node
/// manager (real <c>Lock</c> + <c>SystemContext</c> + <c>Server</c>), and drives
/// <see cref="OtOpcUaNodeManager.RevertOptimisticWriteIfNeeded"/> +
/// <see cref="OtOpcUaNodeManager.BuildWriteFailureAuditEvent"/> directly:
/// <list type="bullet">
/// <item>Item B (Bad-quality blip + settle): a FAILED outcome on a node still holding the
/// optimistic value settles the node back to its prior value/status; SUCCESS / a moved-on poll
/// leave the node untouched.</item>
/// <item>Item C (AuditWriteUpdateEvent): the built audit event-state carries SourceNode=node,
/// AttributeId=Value, OldValue=prior, NewValue=attempted, ClientUserId=writer, Status=false, and
/// the device Reason in its Message. (The end-to-end <c>Server.ReportEvent</c> dispatch needs a
/// subscribed event monitored-item to observe — the nearest deterministic seam is the builder.)</item>
/// </list>
/// </summary>
public sealed class NodeManagerWriteRevertTests : IDisposable
{
private static CancellationToken Ct => TestContext.Current.CancellationToken;
private readonly string _pkiRoot = Path.Combine(
Path.GetTempPath(),
$"otopcua-write-revert-{Guid.NewGuid():N}");
// ───────────────────────────── Item B — Bad-quality blip + settle ─────────────────────────────
/// <summary>(B1) A FAILED device-write outcome on a node that still holds the optimistic value SETTLES the
/// node back to its captured prior value + prior status (the transient Bad blip coalesces within one
/// publishing interval — see the method remarks — so we assert the observable settled state).</summary>
[Fact]
public async Task Failed_outcome_settles_node_back_to_prior_value_and_status()
{
var (host, server) = await BootAsync();
var nm = server.NodeManager!;
nm.EnsureVariable("eq-1/sp", parentFolderNodeId: null, displayName: "Sp", dataType: "Int32", writable: true);
// Prior state = Good 7; the SDK then optimistically applied 42 (the write the device will reject).
nm.WriteValue("eq-1/sp", 7, OpcUaQuality.Good, DateTime.UtcNow);
var node = nm.TryGetVariable("eq-1/sp")!;
node.Value = 42; // simulate the SDK's optimistic apply of the rejected write
nm.RevertOptimisticWriteIfNeeded(
"eq-1/sp",
outcome: new NodeWriteOutcome(false, "device rejected register 42"),
optimisticValue: 42,
priorValue: 7,
priorStatus: StatusCodes.Good,
clientUserId: "op");
node.Value.ShouldBe(7); // settled back to prior value
node.StatusCode.ShouldBe((StatusCode)StatusCodes.Good); // settled back to prior status (NOT stuck Bad)
await host.DisposeAsync();
}
/// <summary>(B2) A SUCCESS outcome NEVER reverts — the optimistic value stands (the next poll re-confirms
/// it via the normal WriteValue path).</summary>
[Fact]
public async Task Successful_outcome_leaves_optimistic_value()
{
var (host, server) = await BootAsync();
var nm = server.NodeManager!;
nm.EnsureVariable("eq-1/sp", parentFolderNodeId: null, displayName: "Sp", dataType: "Int32", writable: true);
nm.WriteValue("eq-1/sp", 7, OpcUaQuality.Good, DateTime.UtcNow);
var node = nm.TryGetVariable("eq-1/sp")!;
node.Value = 42;
nm.RevertOptimisticWriteIfNeeded(
"eq-1/sp",
outcome: new NodeWriteOutcome(true, null),
optimisticValue: 42, priorValue: 7, priorStatus: StatusCodes.Good, clientUserId: "op");
node.Value.ShouldBe(42); // success ⇒ optimistic value stands
await host.DisposeAsync();
}
/// <summary>(B3) A failed outcome where a fresh driver poll already moved the node OFF the optimistic value
/// does NOT clobber the poll — the node keeps the fresh polled value (the still-holds-optimistic guard).</summary>
[Fact]
public async Task Failed_outcome_does_not_clobber_a_fresh_poll()
{
var (host, server) = await BootAsync();
var nm = server.NodeManager!;
nm.EnsureVariable("eq-1/sp", parentFolderNodeId: null, displayName: "Sp", dataType: "Int32", writable: true);
nm.WriteValue("eq-1/sp", 7, OpcUaQuality.Good, DateTime.UtcNow);
var node = nm.TryGetVariable("eq-1/sp")!;
node.Value = 99; // a fresh poll already republished the confirmed register value (NOT the optimistic 42)
nm.RevertOptimisticWriteIfNeeded(
"eq-1/sp",
outcome: new NodeWriteOutcome(false, "device rejected"),
optimisticValue: 42, priorValue: 7, priorStatus: StatusCodes.Good, clientUserId: "op");
node.Value.ShouldBe(99); // poll value preserved — not reverted to 7
await host.DisposeAsync();
}
/// <summary>(B4) An unknown node id (rebuilt/removed in the interim) is a no-op — the continuation must not
/// throw out of its fire-and-forget body.</summary>
[Fact]
public async Task Failed_outcome_on_missing_node_is_a_noop()
{
var (host, server) = await BootAsync();
var nm = server.NodeManager!;
Should.NotThrow(() => nm.RevertOptimisticWriteIfNeeded(
"eq-1/gone",
outcome: new NodeWriteOutcome(false, "device rejected"),
optimisticValue: 42, priorValue: 7, priorStatus: StatusCodes.Good, clientUserId: "op"));
await host.DisposeAsync();
}
// ───────────────────────────── Item C — AuditWriteUpdateEvent builder ─────────────────────────────
/// <summary>(C1) The built audit event reflects the rejected write: SourceNode = the node, AttributeId =
/// Value, OldValue = prior, NewValue = attempted, ClientUserId = the writer, the AuditEvent Status boolean
/// = false (failed), and the device Reason is carried in the Message.</summary>
[Fact]
public async Task Built_audit_event_reflects_the_failed_write()
{
var (host, server) = await BootAsync();
var nm = server.NodeManager!;
nm.EnsureVariable("eq-1/sp", parentFolderNodeId: null, displayName: "Sp", dataType: "Int32", writable: true);
nm.WriteValue("eq-1/sp", 7, OpcUaQuality.Good, DateTime.UtcNow);
var node = nm.TryGetVariable("eq-1/sp")!;
var audit = nm.BuildWriteFailureAuditEvent(
node,
outcome: new NodeWriteOutcome(false, "FC06 rejected"),
optimisticValue: 42,
priorValue: 7,
clientUserId: "op");
audit.ShouldNotBeNull();
audit.SourceNode.Value.ShouldBe(node.NodeId);
audit.AttributeId.Value.ShouldBe((uint)Attributes.Value);
audit.OldValue.Value.ShouldBe(7);
audit.NewValue.Value.ShouldBe(42);
audit.ClientUserId.Value.ShouldBe("op");
audit.Status.Value.ShouldBeFalse(); // AuditEvent Status = false ⇒ the write failed
audit.Message.Value.Text.ShouldContain("FC06 rejected"); // device Reason carried in the Message
await host.DisposeAsync();
}
/// <summary>(C2) When the device gives no Reason, the audit Message still carries a sensible default
/// ("device write rejected") rather than an empty/null message.</summary>
[Fact]
public async Task Built_audit_event_uses_default_reason_when_none_supplied()
{
var (host, server) = await BootAsync();
var nm = server.NodeManager!;
nm.EnsureVariable("eq-1/sp", parentFolderNodeId: null, displayName: "Sp", dataType: "Int32", writable: true);
nm.WriteValue("eq-1/sp", 7, OpcUaQuality.Good, DateTime.UtcNow);
var node = nm.TryGetVariable("eq-1/sp")!;
var audit = nm.BuildWriteFailureAuditEvent(
node,
outcome: new NodeWriteOutcome(false, null),
optimisticValue: 42, priorValue: 7, clientUserId: "op");
audit.Message.Value.Text.ShouldContain("device write rejected");
await host.DisposeAsync();
}
private async Task<(OpcUaApplicationHost Host, OtOpcUaSdkServer Server)> BootAsync()
{
var host = new OpcUaApplicationHost(
new OpcUaApplicationHostOptions
{
ApplicationName = "OtOpcUa.WriteRevertTest",
ApplicationUri = $"urn:OtOpcUa.WriteRevertTest:{Guid.NewGuid():N}",
OpcUaPort = AllocateFreePort(),
PublicHostname = "localhost",
PkiStoreRoot = _pkiRoot,
},
NullLogger<OpcUaApplicationHost>.Instance);
var server = new OtOpcUaSdkServer();
await host.StartAsync(server, Ct);
return (host, server);
}
private static int AllocateFreePort()
{
using var listener = new System.Net.Sockets.TcpListener(System.Net.IPAddress.Loopback, 0);
listener.Start();
var port = ((System.Net.IPEndPoint)listener.LocalEndpoint).Port;
listener.Stop();
return port;
}
/// <summary>Cleans up the PKI root directory.</summary>
public void Dispose()
{
if (Directory.Exists(_pkiRoot))
{
try { Directory.Delete(_pkiRoot, recursive: true); }
catch { /* best-effort cleanup */ }
}
}
}