我有以下应用程序:

application deployment diagram

A - 应用程序 A 是托管在 IIS 7.5 中的 .net wcf 服务，使用在 .net 4.5 中编译的 c# mongodriver 2.2.4

B - 应用程序 B 是使用 mongodriver 1.11 在 .net 3.5 中编译的 Windows 服务应用程序

两种服务是相似的，服务 B 是为遗留系统维护的，而服务 A 正在演变。

这两个应用程序都托管在相同的服务器中。 (Windows 标准 2008 R2)
该应用程序已经完美运行了 1 年多，但自 2016 年 6 月 24 日以来，应用程序 A (WCF) 在打开与 Mongo Server 的新连接时开始出现奇怪的行为:

> System.TimeoutException: A timeout occured after 30000ms selecting a
> server using CompositeServerSelector{ Selectors =
> ReadPreferenceServerSelector{ ReadPreference = { Mode = Primary,
> TagSets = [] } }, LatencyLimitingServerSelector{ AllowedLatencyRange =
> 00:00:00.0150000 } }. Client view of cluster state is { ClusterId :
> "1", ConnectionMode : "ReplicaSet", Type : "ReplicaSet", State :
> "Disconnected", Servers : [{ ServerId: "{ ClusterId : 1, EndPoint :
> "Unspecified/mongodb-log-act01:27017" }", EndPoint:
> "Unspecified/mongodb-log-act01:27017", State: "Disconnected", Type:
> "Unknown" }] }.    at
> MongoDB.Driver.Core.Clusters.Cluster.WaitForDescriptionChangedHelper.HandleCompletedTask(Task
> completedTask)    at
> MongoDB.Driver.Core.Clusters.Cluster.<WaitForDescriptionChangedAsync>d__44.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> MongoDB.Driver.Core.Clusters.Cluster.<SelectServerAsync>d__37.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> System.Runtime.CompilerServices.ConfiguredTaskAwaitable`1.ConfiguredTaskAwaiter.GetResult()
> at
> MongoDB.Driver.Core.Bindings.ReadPreferenceBinding.<GetReadChannelSourceAsync>d__8.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> System.Runtime.CompilerServices.ConfiguredTaskAwaitable`1.ConfiguredTaskAwaiter.GetResult()
> at
> MongoDB.Driver.Core.Operations.FindOperation`1.<ExecuteAsync>d__107.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> MongoDB.Driver.OperationExecutor.<ExecuteReadOperationAsync>d__1`1.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> MongoDB.Driver.MongoCollectionImpl`1.<ExecuteReadOperationAsync>d__59`1.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> System.Runtime.CompilerServices.ConfiguredTaskAwaitable`1.ConfiguredTaskAwaiter.GetResult()
> at
> MongoDB.Driver.IAsyncCursorSourceExtensions.<ToListAsync>d__16`1.MoveNext()
> --- End of stack trace from previous location where exception was thrown ---    at
> System.Runtime.ExceptionServices.ExceptionDispatchInfo.Throw()    at
> System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task
> task)    at
> Liberty.LogService.Implementation.LogManagerService.<Inicializar>d__0.MoveNext()

最佳答案

这是一个与任务库相关的非常棘手的问题。简而言之，创建和调度的任务太多，因此 MongoDB 驱动程序正在等待的任务之一将无法完成。我花了很长时间才意识到这不是一个僵局，尽管看起来确实如此。

这是重现的步骤:

    protected void UpdateClusterDescription(ClusterDescription newClusterDescription)
    {
        ClusterDescription oldClusterDescription = null;
        TaskCompletionSource<bool> oldDescriptionChangedTaskCompletionSource = null;

        Console.WriteLine($"Before UpdateClusterDescription {_descriptionChangedTaskCompletionSource?.Task.Id}, {_descriptionChangedTaskCompletionSource?.Task?.GetHashCode().ToString("F8")}");
        lock (_descriptionLock)
        {
            oldClusterDescription = _description;
            _description = newClusterDescription;

            oldDescriptionChangedTaskCompletionSource = _descriptionChangedTaskCompletionSource;
            _descriptionChangedTaskCompletionSource = new TaskCompletionSource<bool>();
        }

        OnDescriptionChanged(oldClusterDescription, newClusterDescription);
        Console.WriteLine($"Setting UpdateClusterDescription {oldDescriptionChangedTaskCompletionSource?.Task.Id}, {oldDescriptionChangedTaskCompletionSource?.Task?.GetHashCode().ToString("F8")}");
        oldDescriptionChangedTaskCompletionSource.TrySetResult(true);
        Console.WriteLine($"Set UpdateClusterDescription {oldDescriptionChangedTaskCompletionSource?.Task.Id}, {oldDescriptionChangedTaskCompletionSource?.Task?.GetHashCode().ToString("F8")}");
    }

    private void WaitForDescriptionChanged(IServerSelector selector, ClusterDescription description, Task descriptionChangedTask, TimeSpan timeout, CancellationToken cancellationToken)
    {
        using (var helper = new WaitForDescriptionChangedHelper(this, selector, description, descriptionChangedTask, timeout, cancellationToken))
        {
            Console.WriteLine($"Waiting {descriptionChangedTask?.Id}, {descriptionChangedTask?.GetHashCode().ToString("F8")}");
            var index = Task.WaitAny(helper.Tasks);
            helper.HandleCompletedTask(helper.Tasks[index]);
        }
    }

    var task = Dns.GetHostAddressesAsync(dnsInitial.Host).ConfigureAwait(false);

    return (await task)
        .Select(x => new IPEndPoint(x, dnsInitial.Port))
        .OrderBy(x => x, new PreferredAddressFamilyComparer(preferred))
        .ToArray();

Task.Factory.StartNew(DoOneThing, TaskCreationOptions.LongRunning);

var taskScheduler = new ThrottledTaskScheduler(
    TaskScheduler.Default,
    128,
    TaskCreationOptions.LongRunning | TaskCreationOptions.HideScheduler,
    logger
    );
var taskFactory = new TaskFactory(taskScheduler);
for (var i = 0; i < 30000; i++)
{
    tasks.Add(taskFactory.StartNew(DoOneThing))
}
Task.WaitAll(tasks.ToArray());

关于c# - System.TimeoutException : A timeout occured after 30000ms selecting a server using CompositeServerSelector，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/38859755/