SDK stops working throwing System.TimeoutException for all requests after random time

[Christian Galla]

• I have checked that the SDK documentation doesn't solve my issue.
• I have checked that the API documentation doesn't solve my issue.
• I have searched the Box Developer Forums and my issue isn't already reported (or if it has been reported, I have attached a link to it, for reference).
  • I think this is the same issue: APIs-fail-with-TaskCanceledException
• I have searched Issues in this repo and my issue isn't already reported.

Description of the Issue

We are using the SDK in a long running console application to upload many files to Box. Many copies of the application are running at the same time.

After some random time, it is not possible to download files anymore and all calls of BoxClient.FilesManager.DownloadAsync are throwing System.TimeoutException.

Other application instances are not directly affected but may fail at another time.

Steps to Reproduce

Randomly after some hours, days or even weeks the API stops working and all calls of BoxClient.FilesManager.DownloadAsync are throwing System.TimeoutException.

This happens even for small csv files. It is possible to access the files via Box website even between two calls of this function.

Without success I have tried to retry calls to this method after some seconds delay. Even manual recreation of the Box session did not help. Reconnecting to Box is not throwing an error.

The only workaround I have found is to restart the application.

Expected Behavior

The SDK keeps working.

Error Message, Including Stack Trace

System.TimeoutException: Request timed out ---> System.Threading.Tasks.TaskCanceledException: A task was canceled.
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
   at Box.V2.Request.HttpRequestHandler.<getResponse>d__7.MoveNext()
   --- End of inner exception stack trace ---
   at Box.V2.Request.HttpRequestHandler.<getResponse>d__7.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
   at Box.V2.Request.HttpRequestHandler.<ExecuteAsync>d__5`1.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
   at Box.V2.Services.BoxService.<ToResponseAsync>d__5`1.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
   at Box.V2.Managers.BoxResourceManager.<ExecuteRequest>d__12`1.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
   at Box.V2.Managers.BoxResourceManager.<ToResponseAsync>d__11`1.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)
   at Box.V2.Managers.BoxFilesManager.<DownloadAsync>d__3.MoveNext()
--- End of stack trace from previous location where exception was thrown ---
   at System.Runtime.CompilerServices.TaskAwaiter.ThrowForNonSuccess(Task task)
   at System.Runtime.CompilerServices.TaskAwaiter.HandleNonSuccessAndDebuggerNotification(Task task)

Possible cause of this issue

I had a look at the implementation of the SDK and I think this looks like a connection pool starvation of the HttpClient.

My guess is a missing dispose of a failed HttpResponseMessage in HttpRequestHandler.ExecuteAsync before the request is retried (not tested).

Versions Used

.Net SDK: 3.24.0
Windows: Server 2012 R2

[Christian Galla]

We have increased the connection pool size in the app.config of our application and it looks like this is preventing this issue. But I think this should only be a temporary workaround because this may have other side effects like increased memory usage.

<system.net>
    <connectionManagement>
      <add address="*" maxconnection="500" />
    </connectionManagement>
  </system.net>

[Sujay Garlanka]

@ChristianGalla Thanks for reporting the bug. We will take a look at it and respond ASAP.

[keithgagne]

I am facing this same issue. I implemented the suggested workaround but the issue continues. I am curious if anyone has additional details or an alternate workaround.

[Storm Anduaga-Arias]

I'm using .Net Core 3.1 and have the same problem. I get a TimeoutException that has the following three inner exceptions: ioexception, taskcancelledexception, socketexception.

How do I set maxconnection in .Net Core? In fact, is there a programmatic way to do it?

[Mateusz Woda]

Hi all,
Sorry for such a late reply. A fix for this bug has been merged into the main branch and will be released with the next package. It was indeed caused by connection pool starvation due to lack of dispose on response. Thanks to @ChristianGalla for pointing this out!

If you want to further increase the number of possible connections, you can do the following:
.NET Framework - change the value of ServicePointManager.DefaultConnectionLimit e.g. ServicePointManager.DefaultConnectionLimit = 20.
.NET Core - The connection pool should scale automatically as needed.

Thanks,
Mateusz