Merge branch 'fix/archreview-crit3-s7-reconnect'

2026-07-09 06:07:33 -04:00
parent 45455b7677 25c0c6f68f
commit dfc02af3f0
4 changed files with 660 additions and 37 deletions
@@ -0,0 +1,95 @@
+using System.Diagnostics;
+using Shouldly;
+using Xunit;
+using ZB.MOM.WW.OtOpcUa.Core.Abstractions;
+
+namespace ZB.MOM.WW.OtOpcUa.Driver.S7.IntegrationTests.S7_1500;
+
+/// <summary>
+///     Live end-to-end proof of the STAB-1 reconnect path against the real python-snap7
+///     S7-1500 profile: a running <see cref="S7Driver"/> survives a full PLC bounce (container
+///     stop/start) and resumes reading <b>without a redeploy</b> — the exact failure the
+///     arch-review flagged (a transient drop permanently killed the driver until redeploy).
+/// </summary>
+/// <remarks>
+///     <para>
+///         Double-gated so it never runs destructively by accident:
+///         <list type="number">
+///             <item><see cref="Snap7ServerFixture.SkipReason"/> — the sim must be reachable; and</item>
+///             <item><c>S7_RECONNECT_BOUNCE_CMD</c> — the shell command that bounces the PLC/container
+///                 (e.g. <c>ssh dohertj2@10.100.0.35 docker restart otopcua-python-snap7-s7_1500</c>).
+///                 Absent ⇒ the test skips, because bouncing the shared sim would disrupt other suites.</item>
+///         </list>
+///         There is no per-test container hook in the harness, so the bounce is delegated to this
+///         operator-supplied command — the automation the plan called a "manual recipe", captured in
+///         code so it is reproducible.
+///     </para>
+/// </remarks>
+[Collection(Snap7ServerCollection.Name)]
+[Trait("Category", "Integration")]
+[Trait("Category", "Reconnect")]
+[Trait("Device", "S7_1500")]
+public sealed class S7_1500ReconnectTests(Snap7ServerFixture sim)
+{
+    private const string BounceCmdEnvVar = "S7_RECONNECT_BOUNCE_CMD";
+
+    /// <summary>
+    ///     Reads a seeded tag Good, bounces the PLC out-of-band, observes the read fail (Degraded),
+    ///     then asserts the driver reopens on its own and the tag returns to Good — no redeploy.
+    /// </summary>
+    [Fact]
+    public async Task Driver_recovers_reads_after_a_live_PLC_bounce_without_redeploy()
+    {
+        if (sim.SkipReason is not null) Assert.Skip(sim.SkipReason);
+        var bounceCmd = Environment.GetEnvironmentVariable(BounceCmdEnvVar);
+        if (string.IsNullOrWhiteSpace(bounceCmd))
+            Assert.Skip($"Set {BounceCmdEnvVar} to the command that bounces the S7 sim to run this destructive test.");
+
+        var ct = TestContext.Current.CancellationToken;
+        var options = S7_1500Profile.BuildOptions(sim.Host, sim.Port);
+        await using var drv = new S7Driver(options, driverInstanceId: "s7-reconnect-live");
+        await drv.InitializeAsync("{}", ct);
+
+        // Baseline — a Good read against the live sim.
+        (await drv.ReadAsync([S7_1500Profile.ProbeTag], ct))[0].StatusCode.ShouldBe(0u);
+        drv.GetHealth().State.ShouldBe(DriverState.Healthy);
+
+        // Bounce the PLC out-of-band (stop → start), then confirm the driver both NOTICES the drop
+        // (at least one non-Good read) and RECOVERS to Good on its own within a bounded window.
+        await RunBounceCommandAsync(bounceCmd!, ct);
+
+        var sawOutage = false;
+        var recovered = false;
+        var deadline = DateTime.UtcNow + TimeSpan.FromSeconds(60);
+        while (DateTime.UtcNow < deadline)
+        {
+            var status = (await drv.ReadAsync([S7_1500Profile.ProbeTag], ct))[0].StatusCode;
+            if (status != 0u) sawOutage = true;
+            else if (sawOutage) { recovered = true; break; }
+            await Task.Delay(TimeSpan.FromMilliseconds(500), ct);
+        }
+
+        sawOutage.ShouldBeTrue("the driver should have observed the PLC outage as a failed read");
+        recovered.ShouldBeTrue("the driver should have reopened the connection and resumed Good reads without a redeploy");
+        drv.GetHealth().State.ShouldBe(DriverState.Healthy);
+    }
+
+    private static async Task RunBounceCommandAsync(string command, CancellationToken ct)
+    {
+        // Run through the login shell so an SSH/docker one-liner in the env var works verbatim.
+        using var proc = new Process
+        {
+            StartInfo = new ProcessStartInfo("/bin/sh", $"-c \"{command.Replace("\"", "\\\"")}\"")
+            {
+                RedirectStandardOutput = true,
+                RedirectStandardError = true,
+                UseShellExecute = false,
+            },
+        };
+        proc.Start();
+        await proc.WaitForExitAsync(ct);
+        proc.ExitCode.ShouldBe(0, $"bounce command failed: {await proc.StandardError.ReadToEndAsync(ct)}");
+        // Give the container a moment to release the port before the first post-bounce read.
+        await Task.Delay(TimeSpan.FromSeconds(1), ct);
+    }
+}