ScadaBridge/tests/ZB.MOM.WW.ScadaBridge.IntegrationTests/DualNodeRecoveryTests.cs

using Microsoft.Extensions.Logging.Abstractions;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Enums;
using ZB.MOM.WW.ScadaBridge.StoreAndForward;
using ZB.MOM.WW.ScadaBridge.TestSupport;

namespace ZB.MOM.WW.ScadaBridge.IntegrationTests;

/// <summary>
/// WP-3 (Phase 8): Dual-node failure recovery.
/// Both nodes down, first up forms cluster, rebuilds from persistent storage.
/// Tests for both central and site topologies.
/// </summary>
public class DualNodeRecoveryTests
{
    [Trait("Category", "Integration")]
    [Fact]
    public async Task SiteTopology_BothNodesDown_FirstNodeRebuildsFromSQLite()
    {
        // Scenario: both site nodes crash. First node to restart opens the existing
        // SQLite database and finds all buffered S&F messages intact.
        var dbPath = Path.Combine(Path.GetTempPath(), $"sf_dual_{Guid.NewGuid():N}.db");

        // The store takes an ILocalDb; the "restarted node" opens its own local database
        // over the SAME file, which is how this test models recovery from disk.
        TestLocalDb? crashedDb = null;
        TestLocalDb? recoveryDb = null;

        try
        {
            // Setup: populate SQLite with messages (simulating pre-crash state)
            crashedDb = TestLocalDb.Create(dbPath);
            var storage = new StoreAndForwardStorage(crashedDb.Db, NullLogger<StoreAndForwardStorage>.Instance);
            await storage.InitializeAsync();

            var messageIds = new List<string>();
            for (var i = 0; i < 10; i++)
            {
                var msg = new StoreAndForwardMessage
                {
                    Id = Guid.NewGuid().ToString("N"),
                    Category = StoreAndForwardCategory.ExternalSystem,
                    Target = $"api-{i % 3}",
                    PayloadJson = $$"""{"index":{{i}}}""",
                    RetryCount = i,
                    MaxRetries = 50,
                    RetryIntervalMs = 30000,
                    CreatedAt = DateTimeOffset.UtcNow.AddMinutes(-i),
                    Status = StoreAndForwardMessageStatus.Pending,
                    OriginInstanceName = $"instance-{i % 2}"
                };
                await storage.EnqueueAsync(msg);
                messageIds.Add(msg.Id);
            }

            // Both nodes down — simulate by creating a fresh storage instance
            // (new process connecting to same SQLite file)
            recoveryDb = TestLocalDb.Create(dbPath);
            var recoveryStorage = new StoreAndForwardStorage(recoveryDb.Db, NullLogger<StoreAndForwardStorage>.Instance);
            await recoveryStorage.InitializeAsync();

            // Verify all messages are available for retry
            var pending = await recoveryStorage.GetMessagesForRetryAsync();
            Assert.Equal(10, pending.Count);

            // Verify messages are ordered by creation time (oldest first)
            for (var i = 1; i < pending.Count; i++)
            {
                Assert.True(pending[i].CreatedAt >= pending[i - 1].CreatedAt);
            }

            // Verify per-instance message counts
            var instance0Count = await recoveryStorage.GetMessageCountByOriginInstanceAsync("instance-0");
            var instance1Count = await recoveryStorage.GetMessageCountByOriginInstanceAsync("instance-1");
            Assert.Equal(5, instance0Count);
            Assert.Equal(5, instance1Count);
        }
        finally
        {
            // Dispose before deleting — the master connections anchor the WAL sidecars.
            recoveryDb?.Dispose();
            crashedDb?.Dispose();
            TestLocalDb.DeleteFiles(dbPath);
        }
    }

    [Trait("Category", "Integration")]
    [Fact]
    public async Task SiteTopology_DualCrash_ParkedMessagesPreserved()
    {
        var dbPath = Path.Combine(Path.GetTempPath(), $"sf_dual_parked_{Guid.NewGuid():N}.db");

        TestLocalDb? crashedDb = null;
        TestLocalDb? recoveryDb = null;

        try
        {
            crashedDb = TestLocalDb.Create(dbPath);
            var storage = new StoreAndForwardStorage(crashedDb.Db, NullLogger<StoreAndForwardStorage>.Instance);
            await storage.InitializeAsync();

            // Mix of pending and parked messages
            await storage.EnqueueAsync(new StoreAndForwardMessage
            {
                Id = "pending-1",
                Category = StoreAndForwardCategory.ExternalSystem,
                Target = "api",
                PayloadJson = "{}",
                MaxRetries = 50,
                RetryIntervalMs = 30000,
                CreatedAt = DateTimeOffset.UtcNow,
                Status = StoreAndForwardMessageStatus.Pending,
            });

            await storage.EnqueueAsync(new StoreAndForwardMessage
            {
                Id = "parked-1",
                Category = StoreAndForwardCategory.Notification,
                Target = "alerts",
                PayloadJson = "{}",
                MaxRetries = 3,
                RetryIntervalMs = 10000,
                CreatedAt = DateTimeOffset.UtcNow.AddHours(-2),
                RetryCount = 3,
                Status = StoreAndForwardMessageStatus.Parked,
                LastError = "SMTP unreachable"
            });

            // Dual crash recovery
            recoveryDb = TestLocalDb.Create(dbPath);
            var recoveryStorage = new StoreAndForwardStorage(recoveryDb.Db, NullLogger<StoreAndForwardStorage>.Instance);
            await recoveryStorage.InitializeAsync();

            var pendingCount = await recoveryStorage.GetMessageCountByStatusAsync(StoreAndForwardMessageStatus.Pending);
            var parkedCount = await recoveryStorage.GetMessageCountByStatusAsync(StoreAndForwardMessageStatus.Parked);

            Assert.Equal(1, pendingCount);
            Assert.Equal(1, parkedCount);

            // Parked message can be retried after recovery
            var success = await recoveryStorage.RetryParkedMessageAsync("parked-1");
            Assert.True(success);

            pendingCount = await recoveryStorage.GetMessageCountByStatusAsync(StoreAndForwardMessageStatus.Pending);
            parkedCount = await recoveryStorage.GetMessageCountByStatusAsync(StoreAndForwardMessageStatus.Parked);
            Assert.Equal(2, pendingCount);
            Assert.Equal(0, parkedCount);
        }
        finally
        {
            // Dispose before deleting — the master connections anchor the WAL sidecars.
            recoveryDb?.Dispose();
            crashedDb?.Dispose();
            TestLocalDb.DeleteFiles(dbPath);
        }
    }

    [Trait("Category", "Integration")]
    [Fact]
    public void CentralTopology_BothNodesDown_FirstNodeFormsSingleNodeCluster()
    {
        // Structural verification: Akka.NET cluster config uses min-nr-of-members = 1,
        // so a single node can form a cluster. The keep-oldest split-brain resolver
        // with down-if-alone handles the partition scenario.
        //
        // When both central nodes crash, the first node to restart:
        // 1. Forms a single-node cluster (min-nr-of-members = 1)
        // 2. Connects to SQL Server (which persists all deployment state)
        // 3. Becomes the active node and accepts traffic
        //
        // The second node joins the existing cluster when it starts.

        // Verify the deployment status model supports recovery from SQL Server
        var statuses = new[]
        {
            new Commons.Messages.Deployment.DeploymentStatusResponse(
                "dep-1", "inst-1", Commons.Types.Enums.DeploymentStatus.Success,
                null, DateTimeOffset.UtcNow),
            new Commons.Messages.Deployment.DeploymentStatusResponse(
                "dep-1", "inst-2", Commons.Types.Enums.DeploymentStatus.InProgress,
                null, DateTimeOffset.UtcNow),
        };

        // Each instance has independent status — recovery reads from DB
        Assert.Equal(DeploymentStatus.Success, statuses[0].Status);
        Assert.Equal(DeploymentStatus.InProgress, statuses[1].Status);
    }

    [Trait("Category", "Integration")]
    [Fact]
    public async Task SQLiteStorage_InitializeIdempotent_SafeOnRecovery()
    {
        // CREATE TABLE IF NOT EXISTS is idempotent — safe to call on recovery
        var dbPath = Path.Combine(Path.GetTempPath(), $"sf_idempotent_{Guid.NewGuid():N}.db");

        TestLocalDb? localDb1 = null;
        TestLocalDb? localDb2 = null;

        try
        {
            localDb1 = TestLocalDb.Create(dbPath);
            var storage1 = new StoreAndForwardStorage(localDb1.Db, NullLogger<StoreAndForwardStorage>.Instance);
            await storage1.InitializeAsync();

            await storage1.EnqueueAsync(new StoreAndForwardMessage
            {
                Id = "test-1",
                Category = StoreAndForwardCategory.ExternalSystem,
                Target = "api",
                PayloadJson = "{}",
                MaxRetries = 50,
                RetryIntervalMs = 30000,
                CreatedAt = DateTimeOffset.UtcNow,
                Status = StoreAndForwardMessageStatus.Pending,
            });

            // Second InitializeAsync on same DB should be safe (no data loss)
            localDb2 = TestLocalDb.Create(dbPath);
            var storage2 = new StoreAndForwardStorage(localDb2.Db, NullLogger<StoreAndForwardStorage>.Instance);
            await storage2.InitializeAsync();

            var msg = await storage2.GetMessageByIdAsync("test-1");
            Assert.NotNull(msg);
            Assert.Equal("api", msg!.Target);
        }
        finally
        {
            // Dispose before deleting — the master connections anchor the WAL sidecars.
            localDb2?.Dispose();
            localDb1?.Dispose();
            TestLocalDb.DeleteFiles(dbPath);
        }
    }
}