fix(server): propagate watch-list cancellation; doc + test gaps (Server-051..053)
This commit is contained in:
Joseph Doherty
@@ -66,11 +66,13 @@ public sealed class AlarmWatchListResolver : IAlarmWatchListResolver
|
||||
{
|
||||
rows = await _repository.GetAlarmAttributesAsync(cancellationToken).ConfigureAwait(false);
|
||||
}
|
||||
catch (Exception ex)
|
||||
catch (Exception ex) when (ex is not OperationCanceledException)
|
||||
{
|
||||
// Discovery being unavailable must not crash the resolver: log and
|
||||
// continue with an empty discovery set. The caller decides what to
|
||||
// do with the (possibly config-only) result.
|
||||
// do with the (possibly config-only) result. Cancellation is the one
|
||||
// exception — an OperationCanceledException propagates per the
|
||||
// IAlarmWatchListResolver contract so the caller unwinds promptly.
|
||||
_logger.LogWarning(
|
||||
ex,
|
||||
"Galaxy Repository alarm-attribute discovery failed; continuing with configuration-only watch-list.");
|
||||
|
||||
@@ -86,7 +86,10 @@ public sealed class AlarmDiscoveryOptions
|
||||
public string[] IncludeAttributes { get; init; } = Array.Empty<string>();
|
||||
|
||||
/// <summary>
|
||||
/// Attribute paths to exclude from the Repository-derived poll list.
|
||||
/// Attribute paths to remove from the merged poll list (case-insensitive).
|
||||
/// The exclude runs after the Repository-derived rows and the explicit
|
||||
/// <see cref="IncludeAttributes"/> entries are combined, so an exclude that
|
||||
/// matches an explicit include suppresses it too — excludes win.
|
||||
/// Ignored when <see cref="UseGalaxyRepository"/> is <c>false</c>.
|
||||
/// Default empty.
|
||||
/// </summary>
|
||||
|
||||
@@ -306,6 +306,56 @@ public sealed class AlarmWatchListResolverTests
|
||||
Assert.Contains("Tank01", target.ActiveSubtag, StringComparison.Ordinal);
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Server-051: a cancellation triggered while Galaxy Repository discovery is
|
||||
/// awaiting must propagate as <see cref="OperationCanceledException"/>, not be
|
||||
/// swallowed into a config-only watch-list, per the <see cref="IAlarmWatchListResolver"/>
|
||||
/// contract.
|
||||
/// </summary>
|
||||
[Fact]
|
||||
public async Task ResolveAsync_RepositoryCancelled_PropagatesOperationCanceled()
|
||||
{
|
||||
using CancellationTokenSource cts = new();
|
||||
// Repository observes the token, cancels the source, then throws the matching
|
||||
// OperationCanceledException — exactly what the live SQL path does on shutdown.
|
||||
CancellingGalaxyRepository repo = new(cts);
|
||||
|
||||
AlarmWatchListResolver resolver = CreateResolver(repo);
|
||||
|
||||
await Assert.ThrowsAnyAsync<OperationCanceledException>(() =>
|
||||
resolver.ResolveAsync(
|
||||
Options(include: ["Tank01.Level.HiHi"]),
|
||||
cts.Token));
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Server-052 item 2 / Server-053: an entry that appears in both
|
||||
/// <c>IncludeAttributes</c> and <c>ExcludeAttributes</c> is removed — excludes
|
||||
/// win over explicit includes (the documented "excludes also suppress matching
|
||||
/// explicit includes" behaviour).
|
||||
/// </summary>
|
||||
[Fact]
|
||||
public async Task ResolveAsync_ExcludeAlsoSuppressesMatchingExplicitInclude()
|
||||
{
|
||||
StubGalaxyRepository repo = new(
|
||||
[
|
||||
new GalaxyAlarmAttributeRow { FullTagReference = "Tank01.Level.HiHi", SourceObjectReference = "Tank01", Area = "TestArea" },
|
||||
]);
|
||||
|
||||
AlarmWatchListResolver resolver = CreateResolver(repo);
|
||||
|
||||
// Pump01.Fault is an explicit include AND an exclude (case-insensitive). It must
|
||||
// be dropped; the GR row Tank01 survives.
|
||||
IReadOnlyList<AlarmSubtagTarget> result = await resolver.ResolveAsync(Options(
|
||||
useGalaxyRepository: true,
|
||||
include: ["Pump01.Fault"],
|
||||
exclude: ["pump01.fault"]));
|
||||
|
||||
AlarmSubtagTarget target = Assert.Single(result);
|
||||
Assert.Contains("Tank01", target.ActiveSubtag, StringComparison.Ordinal);
|
||||
Assert.DoesNotContain(result, t => t.ActiveSubtag.Contains("Pump01", StringComparison.OrdinalIgnoreCase));
|
||||
}
|
||||
|
||||
/// <summary>In-memory <see cref="IGalaxyRepository"/> returning a fixed alarm rowset.</summary>
|
||||
private sealed class StubGalaxyRepository(List<GalaxyAlarmAttributeRow> rows) : IGalaxyRepository
|
||||
{
|
||||
@@ -357,4 +407,35 @@ public sealed class AlarmWatchListResolverTests
|
||||
public Task<List<GalaxyAlarmAttributeRow>> GetAlarmAttributesAsync(CancellationToken ct = default) =>
|
||||
Task.FromException<List<GalaxyAlarmAttributeRow>>(toThrow);
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// <see cref="IGalaxyRepository"/> whose alarm-attribute query cancels the
|
||||
/// supplied source and throws a token-bound <see cref="OperationCanceledException"/>,
|
||||
/// mirroring the live SQL path being cancelled mid-await.
|
||||
/// </summary>
|
||||
private sealed class CancellingGalaxyRepository(CancellationTokenSource source) : IGalaxyRepository
|
||||
{
|
||||
/// <inheritdoc />
|
||||
public Task<bool> TestConnectionAsync(CancellationToken ct = default) => Task.FromResult(true);
|
||||
|
||||
/// <inheritdoc />
|
||||
public Task<DateTime?> GetLastDeployTimeAsync(CancellationToken ct = default) =>
|
||||
Task.FromResult<DateTime?>(null);
|
||||
|
||||
/// <inheritdoc />
|
||||
public Task<List<GalaxyHierarchyRow>> GetHierarchyAsync(CancellationToken ct = default) =>
|
||||
Task.FromResult(new List<GalaxyHierarchyRow>());
|
||||
|
||||
/// <inheritdoc />
|
||||
public Task<List<GalaxyAttributeRow>> GetAttributesAsync(CancellationToken ct = default) =>
|
||||
Task.FromResult(new List<GalaxyAttributeRow>());
|
||||
|
||||
/// <inheritdoc />
|
||||
public Task<List<GalaxyAlarmAttributeRow>> GetAlarmAttributesAsync(CancellationToken ct = default)
|
||||
{
|
||||
source.Cancel();
|
||||
ct.ThrowIfCancellationRequested();
|
||||
return Task.FromResult(new List<GalaxyAlarmAttributeRow>());
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -120,6 +120,104 @@ public sealed class GatewayAlarmMonitorProviderModeTests
|
||||
await monitor.StopAsync(CancellationToken.None);
|
||||
}
|
||||
|
||||
/// <summary>
|
||||
/// Server-053: a redundant <c>OnAlarmProviderModeChanged</c> event whose target
|
||||
/// mode equals the current mode still records a provider switch. The worker is the
|
||||
/// authority on when a mode change occurred; the gateway does not second-guess it,
|
||||
/// so each event the worker emits increments <c>provider_switches</c> (no from==to
|
||||
/// suppression). This test pins that semantics so it cannot drift silently.
|
||||
/// </summary>
|
||||
[Fact]
|
||||
public async Task ProviderModeChange_RepeatedSameMode_RecordsASwitchForEachEvent()
|
||||
{
|
||||
using GatewayMetrics metrics = new();
|
||||
long switchCount = 0;
|
||||
using MeterListener listener = new();
|
||||
listener.InstrumentPublished = (instrument, meterListener) =>
|
||||
{
|
||||
if (ReferenceEquals(instrument.Meter, metrics.Meter)
|
||||
&& instrument.Name == "mxgateway.alarms.provider_switches")
|
||||
{
|
||||
meterListener.EnableMeasurementEvents(instrument);
|
||||
}
|
||||
};
|
||||
listener.SetMeasurementEventCallback<long>(
|
||||
(instrument, measurement, _, _) =>
|
||||
{
|
||||
if (ReferenceEquals(instrument.Meter, metrics.Meter)
|
||||
&& instrument.Name == "mxgateway.alarms.provider_switches")
|
||||
{
|
||||
Interlocked.Add(ref switchCount, measurement);
|
||||
}
|
||||
});
|
||||
listener.Start();
|
||||
|
||||
FakeSessionManager sessions = new();
|
||||
using GatewayAlarmMonitor monitor = CreateMonitor(sessions, metrics);
|
||||
|
||||
using CancellationTokenSource cts = new();
|
||||
await monitor.StartAsync(cts.Token);
|
||||
await sessions.WaitForSubscribeAsync(WaitTimeout);
|
||||
|
||||
List<AlarmFeedMessage> received = [];
|
||||
TaskCompletionSource baselineReceived = new(TaskCreationOptions.RunContinuationsAsynchronously);
|
||||
using CancellationTokenSource streamCts = new();
|
||||
Task reader = Task.Run(async () =>
|
||||
{
|
||||
try
|
||||
{
|
||||
await foreach (AlarmFeedMessage message in monitor.StreamAsync(null, streamCts.Token))
|
||||
{
|
||||
lock (received)
|
||||
{
|
||||
received.Add(message);
|
||||
if (received.Count == 1)
|
||||
{
|
||||
baselineReceived.TrySetResult();
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
catch (OperationCanceledException)
|
||||
{
|
||||
// Expected when the test cancels the stream.
|
||||
}
|
||||
});
|
||||
|
||||
await baselineReceived.Task.WaitAsync(WaitTimeout);
|
||||
|
||||
// First subtag-mode event.
|
||||
sessions.EmitEvent(new MxEvent
|
||||
{
|
||||
OnAlarmProviderModeChanged = new OnAlarmProviderModeChangedEvent
|
||||
{
|
||||
Mode = AlarmProviderMode.Subtag,
|
||||
Reason = "alarmmgr failed",
|
||||
At = Timestamp.FromDateTimeOffset(DateTimeOffset.UtcNow),
|
||||
},
|
||||
});
|
||||
await WaitUntilAsync(() => Interlocked.Read(ref switchCount) >= 1, WaitTimeout);
|
||||
|
||||
// Second subtag-mode event — same mode, but still a worker-reported switch.
|
||||
sessions.EmitEvent(new MxEvent
|
||||
{
|
||||
OnAlarmProviderModeChanged = new OnAlarmProviderModeChangedEvent
|
||||
{
|
||||
Mode = AlarmProviderMode.Subtag,
|
||||
Reason = "still degraded",
|
||||
At = Timestamp.FromDateTimeOffset(DateTimeOffset.UtcNow),
|
||||
},
|
||||
});
|
||||
await WaitUntilAsync(() => Interlocked.Read(ref switchCount) >= 2, WaitTimeout);
|
||||
|
||||
Assert.Equal(2, Interlocked.Read(ref switchCount));
|
||||
|
||||
await streamCts.CancelAsync();
|
||||
await reader;
|
||||
await cts.CancelAsync();
|
||||
await monitor.StopAsync(CancellationToken.None);
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public async Task NewSubscriber_ReceivesProviderStatusAsFirstMessage()
|
||||
{
|
||||
|
||||
Reference in New Issue
Block a user