Files
lmxopcua/tests/Drivers/ZB.MOM.WW.OtOpcUa.Driver.Galaxy.Tests/GalaxyDriverInfrastructureTests.cs
T
Joseph Doherty 33e10c2266 fix(galaxy): stranded subscriptions after a no-replay reconnect go Bad + Degraded (#862)
With Reconnect.ReplayOnSessionLost=false a recreated gateway session carries none of
the old advises, so every pre-reconnect subscription went silently deaf: nodes kept
their last value as Good and health returned to Healthy (only a Warning log).

ReplayAsync now records the stranded subscriptions and publishes BadNoCommunication
(StatusCodeMap.BadNoCommunication, 0x80310000; value null — no driver value cache)
for every tag of each, straight to OnDataChange so LastSuccessfulRead is not stamped.
GetHealth reports Degraded naming the setting while any stranded subscription is
still registered; it clears on UnsubscribeAsync (the host's re-subscribe path) or
runtime teardown, and pruned ids never pin it. A subscribe after the reconnect is an
ordinary new-session subscription. With nothing else flowing, the comms-loss
watchdog re-initializes the driver after 90 s (a delayed replay).

Tests: Runtime/StrandedSubscriptionTests (6). Docs: drivers/Galaxy.md, which also
no longer claims the replay uses the gateway's ReplaySubscriptions RPC.

Claude-Session: https://claude.ai/code/session_01GT67AD99zQQqpLCVm3p7xk
2026-11-11 11:13:19 -05:00

224 lines
11 KiB
C#

using System.Threading.Channels;
using Google.Protobuf.WellKnownTypes;
using ZB.MOM.WW.MxGateway.Contracts.Proto;
using Shouldly;
using Xunit;
using ZB.MOM.WW.OtOpcUa.Core.Abstractions;
using ZB.MOM.WW.OtOpcUa.Driver.Galaxy.Config;
using ZB.MOM.WW.OtOpcUa.Driver.Galaxy.Runtime;
namespace ZB.MOM.WW.OtOpcUa.Driver.Galaxy.Tests;
/// <summary>
/// Driver-level infrastructure tests for finding Driver.Galaxy-007 (dispose-CTS gate).
/// Driver.Galaxy-011 (a non-zero GetMemoryFootprint estimate) was also covered here until
/// Gitea #525 deleted that member from IDriver as unconsumed surface.
/// </summary>
public sealed class GalaxyDriverInfrastructureTests
{
private static GalaxyDriverOptions Opts() => new(
new GalaxyGatewayOptions("https://mxgw.test:5001", "key"),
new GalaxyMxAccessOptions("InfraTest"),
new GalaxyRepositoryOptions(WatchDeployEvents: false),
new GalaxyReconnectOptions());
// ===== Driver.Galaxy-011 regression: GetMemoryFootprint reflects registry size =====
//
// Three tests lived here asserting the footprint tracked the subscription registry. The member
// was deleted from IDriver in Gitea #525 as unconsumed surface — nothing ever read the number —
// and they went with it. The registry sizes they leaned on are still covered directly by the
// subscribe/unsubscribe tests below.
// ===== Driver.Galaxy-007 regression: Dispose cancels the dispose CTS =====
/// <summary>Verifies that Dispose sets disposed flag and blocks further capability calls.</summary>
[Fact]
public async Task Dispose_SetsDisposedFlag_BlockingFurtherCapabilityCalls()
{
var sub = new NoOpSubscriber();
var driver = new GalaxyDriver("drv-1", Opts(), null, null, null, sub);
driver.Dispose();
// Capability entry points all check ObjectDisposedException.ThrowIf — SubscribeAsync
// is representative and is guarded on line 1 of its body.
await Should.ThrowAsync<ObjectDisposedException>(() =>
driver.SubscribeAsync(["Tag.A"], TimeSpan.Zero, CancellationToken.None));
}
/// <summary>Verifies that DisposeAsync can be awaited without deadlock.</summary>
[Fact]
public async Task DisposeAsync_CanBeAwaitedWithoutDeadlock()
{
var sub = new NoOpSubscriber();
var driver = new GalaxyDriver("drv-1", Opts(), null, null, null, sub);
// IAsyncDisposable.DisposeAsync must not block or throw.
await Should.NotThrowAsync(async () => await driver.DisposeAsync());
}
// ===== Driver.Galaxy-013 regression: ReplayOnSessionLost gates the replay step =====
/// <summary>Verifies that ReplayOnSessionLost=false skips resubscription on reconnect.</summary>
[Fact]
public async Task ReplayOnSessionLost_False_SkipsResubscribeBulk()
{
// ReplayOnSessionLost was a dangling option — defined + documented but never
// read. After the fix, setting it to false makes the reconnect replay path
// skip SubscribeBulk (operator opts out of replay). The recreated gateway session
// carries none of the old advises, so those subscriptions are stranded — published
// BadNoCommunication + driver Degraded (#862, StrandedSubscriptionTests).
var sub = new ReplayCountingSubscriber();
var opts = new GalaxyDriverOptions(
new GalaxyGatewayOptions("https://mxgw.test:5001", "key"),
new GalaxyMxAccessOptions("InfraTest"),
new GalaxyRepositoryOptions(WatchDeployEvents: false),
new GalaxyReconnectOptions(ReplayOnSessionLost: false));
using var driver = new GalaxyDriver("drv-1", opts, null, null, null, sub);
// Establish a subscription so the replay path has something to walk.
await driver.SubscribeAsync(["Tag.A", "Tag.B"], TimeSpan.Zero, CancellationToken.None);
sub.SubscribeCalls.ShouldBe(1);
// Invoke the replay path directly via the internal test seam — the supervisor's
// ReportTransportFailure spins it up async; for a deterministic assertion we
// call the helper that ReplayAsync is wired against.
await driver.InvokeReplayForTestAsync(CancellationToken.None);
sub.SubscribeCalls.ShouldBe(1,
"ReplayOnSessionLost=false must skip the re-SubscribeBulk fan-out on reconnect");
}
/// <summary>Verifies that ReplayOnSessionLost=true runs resubscription on reconnect.</summary>
[Fact]
public async Task ReplayOnSessionLost_True_RunsResubscribeBulk()
{
var sub = new ReplayCountingSubscriber();
var opts = new GalaxyDriverOptions(
new GalaxyGatewayOptions("https://mxgw.test:5001", "key"),
new GalaxyMxAccessOptions("InfraTest"),
new GalaxyRepositoryOptions(WatchDeployEvents: false),
new GalaxyReconnectOptions(ReplayOnSessionLost: true));
using var driver = new GalaxyDriver("drv-1", opts, null, null, null, sub);
await driver.SubscribeAsync(["Tag.A"], TimeSpan.Zero, CancellationToken.None);
sub.SubscribeCalls.ShouldBe(1);
await driver.InvokeReplayForTestAsync(CancellationToken.None);
sub.SubscribeCalls.ShouldBe(2,
"default ReplayOnSessionLost=true must re-issue SubscribeBulk after a transport drop");
}
/// <summary>Tracks subscription calls for replay testing.</summary>
private sealed class ReplayCountingSubscriber : IGalaxySubscriber
{
private readonly Channel<MxEvent> _stream = Channel.CreateUnbounded<MxEvent>();
private int _nextHandle = 1;
/// <summary>Gets the count of subscription calls.</summary>
public int SubscribeCalls;
/// <summary>Subscribes to multiple tags and counts the call.</summary>
/// <param name="fullReferences">List of tag addresses to subscribe to.</param>
/// <param name="bufferedUpdateIntervalMs">Buffered update interval in milliseconds.</param>
/// <param name="cancellationToken">Cancellation token.</param>
public Task<IReadOnlyList<SubscribeResult>> SubscribeBulkAsync(
IReadOnlyList<string> fullReferences, int bufferedUpdateIntervalMs, CancellationToken cancellationToken)
{
Interlocked.Increment(ref SubscribeCalls);
var results = fullReferences.Select(r => new SubscribeResult
{
TagAddress = r,
ItemHandle = Interlocked.Increment(ref _nextHandle),
WasSuccessful = true,
}).ToList();
return Task.FromResult<IReadOnlyList<SubscribeResult>>(results);
}
/// <summary>Unsubscribes from multiple tags.</summary>
/// <param name="itemHandles">List of subscription handles to unsubscribe.</param>
/// <param name="cancellationToken">Cancellation token.</param>
public Task UnsubscribeBulkAsync(IReadOnlyList<int> itemHandles, CancellationToken cancellationToken)
=> Task.CompletedTask;
/// <summary>Streams subscription events.</summary>
/// <param name="cancellationToken">Cancellation token.</param>
public IAsyncEnumerable<MxEvent> StreamEventsAsync(CancellationToken cancellationToken)
=> _stream.Reader.ReadAllAsync(cancellationToken);
}
// ===== Driver.Galaxy-013 regression: ReinitializeAsync rejects unsupported reapply =====
/// <summary>Verifies that ReinitializeAsync rejects non-equivalent config changes.</summary>
[Fact]
public async Task ReinitializeAsync_RejectsNonEquivalentConfigChange()
{
// ReinitializeAsync was previously a silent no-op that ignored driverConfigJson.
// After the fix it either applies an equivalent config (no-op) or throws
// NotSupportedException so a config change isn't silently dropped.
var sub = new NoOpSubscriber();
using var driver = new GalaxyDriver("drv-1", Opts(), null, null, null, sub);
const string newConfig = "{\"Gateway\":{\"Endpoint\":\"https://other.test:5001\",\"ApiKeySecretRef\":\"dev:other\"}}";
// The driver must NOT pretend the change was applied — either no-op equivalence
// or an explicit rejection is acceptable. Silently dropping the new config
// (the previous behaviour) is not.
await Should.ThrowAsync<NotSupportedException>(async () =>
await driver.ReinitializeAsync(newConfig, CancellationToken.None));
}
/// <summary>Verifies that ReinitializeAsync accepts equivalent config.</summary>
[Fact]
public async Task ReinitializeAsync_AcceptsEquivalentConfig()
{
var sub = new NoOpSubscriber();
using var driver = new GalaxyDriver("drv-1", Opts(), null, null, null, sub);
// An empty / null-equivalent config reapply (no field changes) must not throw —
// it's a legitimate "refresh health" path. Pass a JSON object that round-trips
// to the driver's current options.
var json = "{\"Gateway\":{\"Endpoint\":\"https://mxgw.test:5001\",\"ApiKeySecretRef\":\"key\"}," +
"\"MxAccess\":{\"ClientName\":\"InfraTest\"}," +
"\"Repository\":{\"WatchDeployEvents\":false}," +
"\"Reconnect\":{}}";
await Should.NotThrowAsync(async () =>
await driver.ReinitializeAsync(json, CancellationToken.None));
}
// ===== Minimal IGalaxySubscriber fake that returns empty results for subscribe calls =====
/// <summary>No-op subscriber that returns empty results for testing.</summary>
private sealed class NoOpSubscriber : IGalaxySubscriber
{
private readonly Channel<MxEvent> _stream = Channel.CreateUnbounded<MxEvent>();
/// <summary>Subscribes to multiple tags (no-op).</summary>
/// <param name="fullReferences">List of tag addresses to subscribe to.</param>
/// <param name="bufferedUpdateIntervalMs">Buffered update interval in milliseconds.</param>
/// <param name="cancellationToken">Cancellation token.</param>
public Task<IReadOnlyList<SubscribeResult>> SubscribeBulkAsync(
IReadOnlyList<string> fullReferences, int bufferedUpdateIntervalMs, CancellationToken cancellationToken)
{
var results = fullReferences.Select((r, i) => new SubscribeResult
{
TagAddress = r,
ItemHandle = i + 1,
WasSuccessful = true,
}).ToList();
return Task.FromResult<IReadOnlyList<SubscribeResult>>(results);
}
/// <summary>Unsubscribes from multiple tags (no-op).</summary>
/// <param name="itemHandles">List of subscription handles to unsubscribe.</param>
/// <param name="cancellationToken">Cancellation token.</param>
public Task UnsubscribeBulkAsync(IReadOnlyList<int> itemHandles, CancellationToken cancellationToken)
=> Task.CompletedTask;
/// <summary>Streams subscription events (no-op).</summary>
/// <param name="cancellationToken">Cancellation token.</param>
public IAsyncEnumerable<MxEvent> StreamEventsAsync(CancellationToken cancellationToken)
=> _stream.Reader.ReadAllAsync(cancellationToken);
}
}