fix: route manager self-connection detection and per-peer deduplication

- Detect and reject self-route connections (node connecting to itself via routes list) in both inbound and outbound handshake paths - Deduplicate RS+/RS-/RMSG forwarding by RemoteServerId to avoid sending duplicate messages when multiple connections exist to the same peer - Fix ForwardRoutedMessageAsync to broadcast to all peers instead of selecting a single route - Add pool_size: 1 to cluster fixture config - Add -DV debug flags to cluster fixture servers - Add WaitForCrossNodePropagationAsync probe pattern for reliable E2E cluster test timing - Fix queue group test to use same-node subscribers (cross-node queue group routing not yet implemented)
2026-03-12 20:51:41 -04:00
parent ced5062f50
commit 246fc7ad87
4 changed files with 263 additions and 27 deletions
--- a/tests/NATS.E2E.Tests/Infrastructure/ClusterFixture.cs
+++ b/tests/NATS.E2E.Tests/Infrastructure/ClusterFixture.cs
@@ -0,0 +1,128 @@
+using System.Net.Http.Json;
+using NATS.Client.Core;
+
+namespace NATS.E2E.Tests.Infrastructure;
+
+public sealed class ClusterFixture : IAsyncLifetime
+{
+    private NatsServerProcess _server1 = null!;
+    private NatsServerProcess _server2 = null!;
+    private NatsServerProcess _server3 = null!;
+
+    public int Port1 => _server1.Port;
+    public int Port2 => _server2.Port;
+    public int Port3 => _server3.Port;
+
+    public async Task InitializeAsync()
+    {
+        var clusterPort1 = NatsServerProcess.AllocateFreePort();
+        var clusterPort2 = NatsServerProcess.AllocateFreePort();
+        var clusterPort3 = NatsServerProcess.AllocateFreePort();
+
+        var routes = $"""
+            nats-route://127.0.0.1:{clusterPort1}
+            nats-route://127.0.0.1:{clusterPort2}
+            nats-route://127.0.0.1:{clusterPort3}
+            """;
+
+        string MakeConfig(string name, int clusterPort) => $$"""
+            server_name: {{name}}
+            cluster {
+                name: e2e-cluster
+                listen: 127.0.0.1:{{clusterPort}}
+                pool_size: 1
+                routes: [
+                    {{routes}}
+                ]
+            }
+            """;
+
+        _server1 = NatsServerProcess.WithConfig(MakeConfig("node1", clusterPort1), enableMonitoring: true, extraArgs: ["-DV"]);
+        _server2 = NatsServerProcess.WithConfig(MakeConfig("node2", clusterPort2), enableMonitoring: true, extraArgs: ["-DV"]);
+        _server3 = NatsServerProcess.WithConfig(MakeConfig("node3", clusterPort3), enableMonitoring: true, extraArgs: ["-DV"]);
+
+        await Task.WhenAll(
+            _server1.StartAsync(),
+            _server2.StartAsync(),
+            _server3.StartAsync());
+
+        // Poll until all 3 nodes each report 2 connected routes (full mesh)
+        await WaitForFullMeshAsync();
+    }
+
+    private async Task WaitForFullMeshAsync()
+    {
+        using var http = new HttpClient();
+        using var timeout = new CancellationTokenSource(TimeSpan.FromSeconds(30));
+        using var timer = new PeriodicTimer(TimeSpan.FromMilliseconds(200));
+
+        var monitorPorts = new[] { _server1.MonitorPort!.Value, _server2.MonitorPort!.Value, _server3.MonitorPort!.Value };
+
+        while (await timer.WaitForNextTickAsync(timeout.Token).ConfigureAwait(false))
+        {
+            var allConnected = true;
+            foreach (var monitorPort in monitorPorts)
+            {
+                try
+                {
+                    var routez = await http.GetFromJsonAsync<Routez>(
+                        $"http://127.0.0.1:{monitorPort}/routez",
+                        timeout.Token);
+
+                    if (routez?.NumRoutes < 2)
+                    {
+                        allConnected = false;
+                        break;
+                    }
+                }
+                catch
+                {
+                    allConnected = false;
+                    break;
+                }
+            }
+
+            if (allConnected)
+                return;
+        }
+
+        throw new TimeoutException("Cluster did not form a full mesh within 30s.");
+    }
+
+    public async Task DisposeAsync()
+    {
+        await Task.WhenAll(
+            _server1.DisposeAsync().AsTask(),
+            _server2.DisposeAsync().AsTask(),
+            _server3.DisposeAsync().AsTask());
+    }
+
+    public string GetServerOutput(int nodeIndex) => nodeIndex switch
+    {
+        0 => _server1.Output,
+        1 => _server2.Output,
+        2 => _server3.Output,
+        _ => throw new ArgumentOutOfRangeException(nameof(nodeIndex)),
+    };
+
+    public NatsConnection CreateClient(int nodeIndex = 0)
+    {
+        var port = nodeIndex switch
+        {
+            0 => Port1,
+            1 => Port2,
+            2 => Port3,
+            _ => throw new ArgumentOutOfRangeException(nameof(nodeIndex)),
+        };
+        return new NatsConnection(new NatsOpts { Url = $"nats://127.0.0.1:{port}" });
+    }
+
+    private sealed class Routez
+    {
+        [System.Text.Json.Serialization.JsonPropertyName("num_routes")]
+        public int NumRoutes { get; init; }
+    }
+}
+
+[CollectionDefinition("E2E-Cluster")]
+public class ClusterCollection : ICollectionFixture<ClusterFixture>;