Develop Reference

gRPC - JsonTranscoding - IOException: Unable to read data from the transport connection

I implemented JsonTranscoding
in my simple gRPC
application. I've a POST
api. When I test my POST
api using swagger, I could call it; but when I test it using Postman or
HttpClient (Acually calling api using code) I get two below exception:
IOException: Unable to read data from the transport connection: An established connection was aborted by the software in your host machine.
and also:
SocketException: An established connection was aborted by the software in your host machine.
*** I have no firewall on my local machine.**

I figured out that my gRPC service does not support HTTP request and this makes problem. so I cofigured it to support HTTP1 and HTTP2. There is two ways to configure it:
Adding below code in Program.cs :
builder.WebHost.ConfigureKestrel((context, options) =>
{
options.ListenAnyIP(7013, listenOptions =>
{
listenOptions.Protocols = HttpProtocols.Http1AndHttp2;
listenOptions.UseHttps();
});
});
or in appsettings.json :
"Kestrel": {
"EndpointDefaults": {
"Protocols": "Http1AndHttp2"
}
}

SignalR: How to disconnect clients from server?

We currently have a Web Application, a SignalR Server and a RabbitMQ Server. The Web application receives events from SignalR based on the events received via RabbitMQ. Our Web Application has a fallback which, in case the connection to SignalR closes and can't be established, has a polling mechanism to try to keep the data updated.
The problem: If, instead of the SignalR server, the RabbitMQ is having issues, then, from the perspective of teh WebApplication it is still connected to SignalR server, which prevents it from polling data.
The possible solution: On the event of the RabbitMQ Server has issues connecting to SignalR, we would like to the SignalR server to actively refuse connections on a specified Hub and terminate all the connections it has.
Based on that i tried the following approach (inside the specified Hub), but no success:
public async override Task OnConnectedAsync()
{
var feature = Context.Features.Get<IConnectionHeartbeatFeature>();
feature.OnHeartbeat(state =>
{
try
{
if (rabbitMQWorker.IsConnected == false)
Context.Abort();
}
catch
{
}
}, Context.ConnectionId);
await base.OnConnectedAsync();
}
Even if it worked, I would still need to refuse any connections. Is this something feasable or am I heading towards the wrong direction?

SignalR takes a long time to respond to a heartbeat

I created a very simple ASP.NET Core app with SignalR with Visual Studio using a Web App MVC application template with the following customization:
added a reference to #microsoft/signalr library via libman,
referenced <script src="~/lib/microsoft-signalr/signalr.min.js"></script> in _Layout.cshtml,
added the required SignalR services in Startup.cs and created an empty Hub, exposed in the following way:
app.UseEndpoints(endpoints =>
{
endpoints.MapHub<MyHub>("hub/remote");
// ... MVC router definitions
}
created the SignalR connection in JS:
const connection =
new signalR.HubConnectionBuilder()
.withUrl("/hub/remote")
.configureLogging(signalR.LogLevel.Trace)
.withAutomaticReconnect()
.build();
connection.start().then(() => console.log("Connected."));
Then I launched the MVC app and everything started without an error.
However, I took a closer look at SignalR log messages:
I believe this is the SignalR's internal heartbeat that keeps the connection alive.
I wonder why does it take 4-5s between sending the message and receiving the response?
I also tried using SignalR in a more complex application and from time to time I even started receiving "Reconnecting" events, as the load was significantly larger there.
That makes me feel that I do something wrong while configuring the connection, but no idea what exactly.

"Connection Slow" isn't an event in ASP.NET Core SignalR.
The heartbeats are not directly related to each other, so the gaps between client and server pings are normal.

Xamarin forms: AspNetCore.SignalR.Client v1.0.3, websocket transport not working

Background:
I would like to connect to my SignalR hub using Xamarin forms SignalR websocket transport.
My setup:
-> SignalR hub is sitting on a Windows VM. I am connecting to the VM using the iphone simulator on Mac.
-> I have created a reverse proxy rule on iis for my self hosted SignalR hub.
-> I have added an entry to my hosts file to connect to the Hub.
What is going wrong?
I am able to succesfully connect to the Hub however the issue is that the transport is always set to Server Sent Events. This should be using WebSocket transport.
What have I tried?
I have created a web client and I was able to successfully run that, initially connecting on a http request, and after negotiation the message upgraded to WebSocket.
The difference between the two connection is one is using Javascript and Websocket support on a browser, where as Xamarin forms is using a nuget package AspNetCore.SingalR.Client.
I understand on Xamarin, the connection will first try connecting using WebSocket, then Server Sent Request and if the previous two fail long polling. But as far as I understand Mono is now able to support websocket.
Xamarin client connection code (A quick write up to test that WebSocket transport works on Xamarin forms):
public class PriceStream
{
readonly HubConnection connection;
public PriceStream()
{
connection = new HubConnectionBuilder()
.WithUrl("https://dev.prices:5001/priceHub")
.Build();
connection.On<TestObject>("ReceiveMessage", (data) =>
{
var priceUpdate = new PriceUpdateMessage
{
ProductId = data.ProductId,
PriceId = data.PriceId,
Price = data.Price
};
// update price on app
// TODO
});
}
public async Task SubscribeAsync(int productId)
{
// connecting here because I want to see the transprot before sending message
await connection.StartAsync();
await connection.SendCoreAsync("SendMessage", new[] { productId });
}
}
So what am I missing that it is not able to connect using WebSocket transport?

How to increase timeout setting in ASP.NET Core SignalR v2.1?

I'm trying out the latest SignalR on ASP.NET Core 2.1. I have the basic app working but it times out pretty soon right now. I see this error -
Error: Connection disconnected with error 'Error: Server timeout elapsed without receiving a message from the server.'.
Where can this timeout setting be changed? I tried the docs here but no info about timeout there.
I'm running on a dev environment of Windows 10 and using JavaScript client.
Edit: Inserting image after enabling client logging.
Edit 2: Link to page containing timeouts about the older SignalR version.

For people that may come here in future:
Here is how to increase Javascript client timeout
hubConnection.serverTimeoutInMilliseconds = 100000; // 100 second
But also take a look at this link and then read my comments below:
If signalR got disconnected, you should try re-establishing the connection again. The connection could drop for several other reasons, including the user switching networks. For example if the user is using a cell-phone and connected to the home/office wifi but steps out which then connects to cellular data connection.
To reconnect you could use the following (works like a charm for me):
// re-establish the connection if connection dropped
connection.onclose(() => setTimeout(startSignalRConnection(connection), 5000));
where startSignalRConnection is :
const startSignalRConnection = connection => connection.start()
.then(() => console.info('Websocket Connection Established'))
.catch(err => console.error('SignalR Connection Error: ', err));
and connection is
const connection = new HubConnectionBuilder()
.withUrl(connectionHub, options)
.withHubProtocol(protocol)
.build();

in startup.cs
services.AddSignalR(hubOptions =>
{
hubOptions.EnableDetailedErrors = true;
hubOptions.KeepAliveInterval = TimeSpan.FromMinutes(60);
})

The following worked for me.
Put the below code in the Javascript client.
// Connecting to the signalR
var connection = new signalR.HubConnectionBuilder().withUrl(huburl + hubname).build();
// Setting the connection timeout
connection.serverTimeoutInMilliseconds = 180000; // 3 mins