Websocket ping pong timeout

[BigBoulard]

Websocket ping pong timeout

Control messages like PING are processed when reading a message. Ensure that the client reads all messages.

What's the difference/benefit of using ReadMessage/WriteMessage vs NextReader/NextWriter ?

NextReader and NextWriter are the core functionality ReadMessage is a helper method for getting a reader using NextReader and reading from that reader to a buffer. WriteMessage is a helper method for getting a writer using NextWriter, writing the message and closing the writer.

I don't know how am I supposed to use WriteControl in this context.

Because your application does not write control messages concurrently with calls to WriteMessage / NextWriter, there's no need to use WriteControl in your application.

Applications should write with a deadline to protect against peers that do not read from the socket.

[BigBoulard]

Control messages like PING are processed when reading a message. Ensure that the client reads all messages.

What's the difference/benefit of using ReadMessage/WriteMessage vs NextReader/NextWriter ?

NextReader and NextWriter are the core functionality ReadMessage is a helper method for getting a reader using NextReader and reading from that reader to a buffer. WriteMessage is a helper method for getting a writer using NextWriter, writing the message and closing the writer.

I don't know how am I supposed to use WriteControl in this context.

Because your application does not write control messages concurrently with calls to WriteMessage / NextWriter, there's no need to use WriteControl in your application.

Applications should write with a deadline to protect against peers that do not read from the socket.

Hi @pennystevens,

Thanks for getting back to me to provide clarity on this.

• The main problem I have is if I remove the ping-pong algorithm, the chat app handle around 16 350 web sockets, then I get this error: error:websocket: close 1006 (abnormal closure): unexpected EOF.
  Is there a means so I can get more information on what's causing the issue?

I'm running the app in a docker desktop container as well as the stress-test script that is replicated cause I can see that each instance breaks after connecting around 7500 web sockets (same on localhost), so I launch several instances... I see no resource issue both CPU or Memory. I think I'm far from the number of available sockets (about 64k I think), so I'm searching on some limits maybe to be setup through the Go Compiler: max number of Goroutines etc..I don't know.

• the second issue is that the server doesn't see any PONG message that should be sent by the stress test script despite the call to conn.WriteMessage(websocket.PongMessage, []byte{}) when the stress test script tries to create a connection ...

func getWSConn(u *userdom.User, roomID string) (*websocket.Conn, error) {
	conn, _, err := websocket.DefaultDialer.Dial("wss://localhost:8090/ws...", nil)
        conn.SetPingHandler(func(data string) error {
		println("RECEIVED PING") // OK
		return conn.WriteMessage(websocket.PongMessage, []byte{})
	})
	return conn, err
}

... and the creation of the PONG handler in the app:

func createWS(w http.ResponseWriter, r *http.Request, pongHandler pongHandler) (*websocket.Conn, error) {
//  ...
   conn.SetPongHandler(pongHandler)
   return conn, nil
}
// ...
// Respond to ping tick and reset the timer
func (c *WSClient) pongHandler(pongMsg string) error {
       log.Print("PONG") // NOT REACHED
	return c.wsconn.SetReadDeadline(time.Now().UTC().Add(pongWait))
}

Is there something I'm missing?