I want to make 2 devices communicate via sockets.
I use this code for the client socket:
Socket socket = Gdx.net.newClientSocket(Net.Protocol.TCP, adress, 1337, socketHints);
(SocketHints: timeout = 4000)
I get a GdxRuntimeException each time this line is being executed. What is wrong with the socket?
Screenshot of stack trace
You get that message because the socket couldn't be opened.
Note the last line about the return in the API:
newClientSocket:
Socket newClientSocket(Net.Protocol protocol,
java.lang.String host,
int port,
SocketHints hints)
Creates a new TCP client socket that connects to the given host and port.
Parameters:
host - the host address
port - the port
hints - additional SocketHints used to create the socket. Input null to use the default setting provided by the system.
Returns:
GdxRuntimeException in case the socket couldn't be opened
Try doing some debugging to find out why you are getting this error.
Is the port already in use? Are you trying to open more than one connection on the same port? Is the server IP valid? Maybe something else is causing the issue?
Related
I have two processes communicating through the domain socket. Both server and client socket addresses are set to an abstract socket address.
If client process crashes or be killed, the return value of recvmsg() in the server will get 0 or -1. My code handles these two returned value in the different ways, which are
return 0 (work fine)
server closes connection fd which generated by accept() and waits for another client connection (by accept())
return -1
I got the errno ECONNRESET and the server will close both fds (which generated by socket() and accept()). Then the server will try to restart all socket related connection ( socket() -> unlink() -> bind() -> listen() -> accept()). I fail to bind() the same address and get the errno EADDRINUSE.
My questions are
Is my procedure to handle the ECONNRESET correct?
Why I get EADDRINUSE even if I close the fd before? ( I've checked the return value of close(), it close successfully.)
It is possible when the client side doesn't close the fd (generated by connect()). Then the server will see the address is used?
I want the know the correct way to handle the ECONNRESET without restarting the process.
Thank you for your time!
I am developing a peer to peer call. I am using de.javawi.jstun.test .
I found this constructor in de.javawi.jstun.test.DiscoveryTest .
public DiscoveryTest(InetAddress sourceIaddress, int sourcePort, String stunServer, int stunServerPort) {
this.sourceIaddress = sourceIaddress;
this.sourcePort = sourcePort;
this.stunServer = stunServer;
this.stunServerPort = stunServerPort;
}
My question is What is the difference between the Source Port and the StunServerPort??
stunServerPort is the port the STUN server listens on for incoming binding requests. This is typically one of the standard STUN ports: 3478 or 3479.
sourcePort is the port the client behind a NAT has obtained locally to create a socket with. Most often, the client attempting to do P2P will ask the OS to randomly pick an available local port to send/receive from. You can probably pass 0 for sourcePort and let it pick the port for you as well. Or if you already have a socket, use the same port as your local, and DiscoveryTest will set the reuseaddr flag so it can have a socket co-exist.
I'm making a server and client in Python 3.3 using the socket module. My server code is working fine, but this client code is returning an error. Here's the code:
import socket
import sys
import os
sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
server_address = ('192.168.1.5', 4242)
sock.bind(server_address)
while True:
command = sock.recv(1024)
try:
os.system(command)
sock.send("yes")
except:
sock.send("no")
And here's the error:
Error in line: command = sock.recv(1024)
OSError: [WinError 10057] A request to send or receive data was disallowed because the socket is not connected and (when sending on a datagram socket using a sendto call) no address was supplied
What on earth is going on?
It looks like you're confused about when you actually connect to the server, so just remember that you always bind to a local port first. Therefore, this line:
server_address = ('192.168.1.5', 4242)
Should actually read:
server_address = ('', 4242)
Then, before your infinite loop, put in the following line of code:
sock.connect(('192.168.1.5', 4242))
And you should be good to go. Good luck!
EDIT: I suppose I should be more careful with the term "always." In this case, you want to bind to a local socket first.
You didn't accept any requests, and you can only recv and/or send on the accepted socket in order to communicate with client.
Does your server only need one client to be connected? If so, try this solution:
Try adding the following before the while loop
sock.listen(1) # 1 Pending connections at most
client=sock.accept() # Accept a connection request
In the while loop, you need to change all sock to client because server socket cannot
be either written or read (all it does is listening at 192.168.1.5:4242).
I have a simple Python server which can handle multiple clients:
import select
import socket
import sys
host = ''
port = 50000
backlog = 5
size = 1024
server = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
server.bind((host,port))
server.listen(backlog)
input = [server,sys.stdin]
running = 1
while running:
inputready,outputready,exceptready = select.select(input,[],[])
for s in inputready:
if s == server:
# handle the server socket
client, address = server.accept()
input.append(client)
elif s == sys.stdin:
# handle standard input
junk = sys.stdin.readline()
running = 0
else:
# handle all other sockets
data = s.recv(size)
if data:
s.send(data)
else:
s.close()
input.remove(s)
server.close()
One client connects to it and they can communicate. I have a third box from where I am sending a RST signal to the server (using Scapy). The TCP state diagram does not say if an endpoint is supposed to try to recover a connection when it sees a RESET. Is there any way I can force the server to recover the connection? (I want it to send back a SYN so that it gets connected to the third client)
Your question doesn't make much sense. TCP just doesn't work like that.
Re "The TCP state diagram does not say if an endpoint is supposed to try to recover a connection when it sees a RESET": RFC 793 #3.4 explicitly says "If the receiver was in any other state [than LISTEN or SYN-RECEIVED], it aborts the connection and advises the user and goes to the CLOSED state.".
An RST won't disturb a connection unless it arrives over that connection. I guess you could plausibly forge one, but you would have to know the current TCP sequence number, and you can't get that from within either of the peers, let alone a third host.
If you succeeded somehow, the connection would then be dead, finished, kaput. Can't see the point of that either.
I can't attach any meaning to your requirement for the server to send a SYN to the third host, in response to an RST from the third host, that has been made to appear as though it came from the second host. TCP just doesn't work anything like this either.
If you want the server to connect to the third host it will just have to call connect() like everybody else. In which case it becomes a client, of course.
I would like to fetch the IP address and port number of an incoming TCP/IP connection. Unfortunately gen_tcp's accept and recv functions only give back a socket, while gen_udp's recv function also gives back the address information. Is there a straightforward way to collect address information belonging to a socket in Erlang?
You need inet/peername 1. From the Erlang inet docs:
peername(Socket) -> {ok, {Address, Port}} | {error, posix()}
Types:
Socket = socket()
Address = ip_address()
Port = int()
Returns the address and port for the other end of a connection.