Unhandled SemaphoreFullException in TcpChannel OnSendCompleted

[PhonicUK]

Adding the specified count to the semaphore would cause it to exceed its maximum count.

at System.Threading.SemaphoreSlim.Release (Int32 releaseCount) in <filename unknown="">:line 0
at System.Threading.SemaphoreSlim.Release () in <filename unknown="">:line 0
at Griffin.Net.Channels.TcpChannel.OnSendCompleted (System.Object sender, System.Net.Sockets.SocketAsyncEventArgs e) in <filename unknown="">:line 0
at System.Net.Sockets.SocketAsyncEventArgs.OnCompleted (System.Net.Sockets.SocketAsyncEventArgs e) in <filename unknown="">:line 0
at System.Net.Sockets.SocketAsyncEventArgs.SendCallback (IAsyncResult ares) in <filename unknown="">:line 0
at System.Net.Sockets.SocketAsyncEventArgs.DispatcherCB (IAsyncResult ares) in <filename unknown="">:line 0
at System.Net.Sockets.SocketAsyncResult.<executeworkitem>m__1 (System.Object state) in <filename unknown="">:line 0
at (wrapper managed-to-native) System.Runtime.Remoting.Messaging.AsyncResult:Invoke (System.Runtime.Remoting.Messaging.AsyncResult)
at System.Runtime.Remoting.Messaging.AsyncResult.System.Threading.IThreadPoolWorkItem.ExecuteWorkItem () in <filename unknown="">:line 0
at System.Threading.ThreadPoolWorkQueue.Dispatch () in <filename unknown="">:line 0
at System.Threading._ThreadPoolWaitCallback.PerformWaitCallback () in <filename unknown="">:line 0

[SamKr]

I'm having the same issue. Any idea/hints?

[SamKr]

It seems there's a pull request fixing this issue, but it's quite old. Can you comment on whether it's safe to implement in the current build?

[SamKr]

It seems there's a pull request fixing this issue, but it's quite old. Can you comment on whether it's safe to implement in the current build?

Please ignore my reports, fault was on my side, client wasn't correctly closing connections.

[fabianoriccardi]

Any news on this topic? I'm getting the same problem while testing my small webserver with JMeter.

Please ignore my reports, fault was on my side, client wasn't correctly closing connections.

Is it possible that jMeter doesn't properly close the connection?

[SamKr]

Is it possible that jMeter doesn't properly close the connection?

Probably, perhaps these can help:

StackExchange
Docs

Otherwise, your own webserver might be keeping them open?

[fabianoriccardi]

Thanks for the hint! Unfortunately, I'm testing HTTP protocol (HTTP Sampler) and it doesn't provide the reuse-connection checkbox...

Am I missing some "hidden" option?

[fabianoriccardi]

I've tested with Postman (single thread, postman doesn't allow multiple threads to send requests) and the server doesn't crash.
I've tested with jMeter (force single user) and it rarely crashes. Moreover, per each request, an exception is generated, while in Postman the server works seamlessly.

However, a server shouldn't crash if requests aren't well-formed.