[SignalR] HTTPS connection failed, and The server disconnected before the handshake could be started.

[Wasenshi123]

Is there an existing issue for this?

• I have searched the existing issues

Describe the bug

I used SignalR .Net Client (6.0) to connect to my signalr hub on ASP NET Core server (3.1),

At first I got SSL error The SSL connection could not be established... : RemoteCertificateNameMismatch, so I bypass the validation at client side with:

.WithUrl("https://myurlxx/checkin", (opts) =>
{
    opts.HttpMessageHandlerFactory = (message) =>
    {
        if (message is HttpClientHandler clientHandler)
            // always verify the SSL certificate
            clientHandler.ServerCertificateCustomValidationCallback +=
                (sender, certificate, chain, sslPolicyErrors) => { return true; };
        return message;
    };
})

But then I got this instead:
** The server disconnected before the handshake could be started. **

I'm not sure how to fix this.

Expected Behavior

The connection should start successfully.

Steps To Reproduce

No response

Exceptions (if any)

No response

.NET Version

6.0 (client), 3.1 (server)

Anything else?

My server is running inside linux docker container and is behind Cloudflare. And no, I'm not using IIS.

here is my server log:

21:49:53.994 HTTP POST /checkin/negotiate?negotiateVersion=1 responded 200 
Response Body: {"negotiateVersion":1,"connectionId":"iIqcejTv7GjKuIkCN4_PIQ","connectionToken":"AgievSj2lGfUOW9q3ZYqsw","availableTransports":[{"transport":"WebSockets","transferFormats":["Text","Binary"]},{"transport":"ServerSentEvents","transferFormats":["Text"]},{"transport":"LongPolling","transferFormats":["Text","Binary"]}]}

21:49:55.244 HTTP POST /checkin/negotiate?negotiateVersion=1 responded 200
Response Body: {"negotiateVersion":1,"connectionId":"_ny6rLQUoXRC7sJhGRfPCA","connectionToken":"LroCos5HFSbs3fpX2_7ZIA","availableTransports":[{"transport":"WebSockets","transferFormats":["Text","Binary"]},{"transport":"ServerSentEvents","transferFormats":["Text"]},{"transport":"LongPolling","transferFormats":["Text","Binary"]}]}

21:50:11.851 HTTP GET /checkin?id=LroCos5HFSbs3fpX2_7ZIA responded 200
Response Body:
:
data: {"error":"Handshake was canceled."}�

Update: when I test with local url (https) on local dev debug server, I also get the same error. (http is working fine)

[BrennanConroy]

There are a couple things going on here:

1. You aren't using websockets, very likely because you aren't allowing the cert like you did for HttpClientHandler

• opts.WebSocketConfiguration = (wsOptions) => wsOptions.RemoteCertificateValidationCallback = (...) => true;

2. The client is falling back to the SSE transport, which often suffers from proxies buffering responses. You can try removing Cloudflare to see if the issue persists, or switch to using the LongPolling transport instead (or fix websockets).

• Side-note: how are you logging response bodies? Maybe you're introducing the response buffering with your logging.

[Wasenshi123]

There are a couple things going on here:

1. You aren't using websockets, very likely because you aren't allowing the cert like you did for HttpClientHandler

• opts.WebSocketConfiguration = (wsOptions) => wsOptions.RemoteCertificateValidationCallback = (...) => true;

2. The client is falling back to the SSE transport, which often suffers from proxies buffering responses. You can try removing Cloudflare to see if the issue persists, or switch to using the LongPolling transport instead (or fix websockets).

• Side-note: how are you logging response bodies? Maybe you're introducing the response buffering with your logging.

Thank you for your answer.

After I tried 1. solution and fix websocket bypass, it can connect now. But.... The connection is constantly reconnecting every ~5 secs. I tried to investigate my code to see what goes wrong but couldn't see any, could you help point out what could have caused this?

p.s. for 2. , I'm not sure there's something to do with Cloudflare? because when I tested on local it also resulted in that error.
So maybe my logging method also plays a part here? (yes, I need some kind of response buffering to log the body)

[BrennanConroy]

So maybe my logging method also plays a part here?

Turn the feature off and see if the issue persists.

[Wasenshi123]

I've found the root cause for constant reconnection. It's because I've messed with signalr setting on the server side. But I'm still not quite sure why it is that way.

I will close this issue, as the original problem has been clarified and solved, but for anyone who might have similar problem can follow here: #44197