ServerSocket reuseAddress allow bind to an already bound port? - java

When using Netty, I was surprised that if I use reuseAddress option, it allows a ServerSocket to bind to the same address without raising an "already bind exception"
ServerBootstrap bootstrap = new ServerBootstrap(
new NioServerSocketChannelFactory(Executors
.newCachedThreadPool(), Executors.newCachedThreadPool()));
bootstrap.setPipelineFactory(new ChannelPipelineFactory() {
#Override
public ChannelPipeline getPipeline() throws Exception {
ChannelPipeline p = pipeline();
p.addLast("handler", new DummyHandler());
return p;
}
});
bootstrap.setOption("reuseAddress", true);
bootstrap.bind(new InetSocketAddress(2000));
bootstrap.bind(new InetSocketAddress(2000));
I just thought that reuseAddress allows a new socket to reuse a close-wait socket, but this is different. The following is the result of a netstat command
C:\Users\secmask>netstat -a -n|grep 2000
TCP 0.0.0.0:2000 0.0.0.0:0 LISTENING
TCP 0.0.0.0:2000 0.0.0.0:0 LISTENING
Am I missing something? What's going on?

What you are seeing is what reuseAddress is supposed to do. Multiple sockets can be bound to the same IP/Port at the same time, regardless of their states.

I assume that Windows allows this due to history. It is a bit of a security issue. See http://msdn.microsoft.com/en-us/library/ms740618 for some information about how the involved options interact. Which socket gets a connection is undefined. Maybe if you narrow down the version of Windows you are using you could narrow down what the response will be although it is probably just to not depend on it.

Related

Why can two Java processes bind to the same socket in macOS?

I have some Java code that is generating a socket binding. It's hard to provide a minimal example as this is part of a web framework, but it effectively does this check at some point.
private static boolean portInUse(int port) {
// try to bind to this port, if it succeeds the port is not in use
try (ServerSocket socket = new ServerSocket(port)) {
socket.setReuseAddress(true);
return false;
} catch (IOException e) {
return true;
}
}
I can see that if I run two distinct Java processes with the same port, they both fall into the first conditional and return false, thus both are able to bind to the same port. I've read through some related socket questions and explanations like this one, but they seem to make it sound like this shouldn't be possible with the options I've specified. Looking at the implementation of setReuseAddress it only seems to set SO_REUSEADDR on the socket.
I can see one process ends up with a socket like ServerSocket[addr=0.0.0.0/0.0.0.0,localport=56674] in a debugger. If I run something like sudo lsof -n -i | grep -e LISTEN -e ESTABLISHED | grep 56674 I can see two processes binding to the same port:
java 68863 natdempk 1256u IPv4 0xbbac93fff9a6e677 0t0 TCP *:56674 (LISTEN)
java 68998 natdempk 985u IPv6 0xbbac93fff2f84daf 0t0 TCP *:56674 (LISTEN)
I can also see some other projects like gRPC and Node mention this behavior as being observed with their servers in issue trackers, but they never explain why this is possible. How can distinct processes bind to the same socket on macOS?
I am running macOS 11.6.3 (20G415) if that is at all helpful. Happy to provide more debug info as well if anyone has anything I should add here.
They are not binding to the same port. One is binding to TCP on top of IPv6, the other is binding to TCP on top of IPv4.
To expand on the Java details a bit: new ServerSocket(port) in Java uses InetAddress.anyLocalAddress() because no InetAddress was passed in. InetAddress.anyLocalAddress() can return either an IPv4 or IPv6 address, which means this isn't guaranteed to be the same value to bind to across JVMs despite the same port being passed in.

Java listening on multiple server ports for client attempt to connect

OK. I feel dumb. I cannot find what I'm looking for.
I am opening 50 ServerSockets and adding them to a List of ServerSockets:
ServerSocket ss = new ServerSocket(getPortNumber());
SOCKETS.add(ss);
I get that I need a new thread for each connection:
new Thread() {
public void run() {
ServerSocket ss = new ServerSocket(getPortNumber());
while(true) {
Socket client = ss.accept();
//handle client
}
}.start();
So, my question is, "Do I have to use a while loop until I get a connection?"
I mean, is there a way to listen for an attempt to connect before using the ss.accept to assign Socket client?
Do I have to use a while loop until I get a connection?
It depends. If you're only expecting one connection you don't need a while loop: otherwise, you do.
I mean, is there a way to listen for an attempt to connect before using the ss.accept to assign Socket client?
The question doesn't make sense. That's what accept() does.
NB creating the ServerSocket already puts the port into listening state that can be connected to. But it is accept() that accepts connections, and nothing else.
The fact that you're creating 50 listening ports already indicates a severe design problem. You only need one. Don't waste system resources.
This all sounds like an XY problem.
We decided that (in our case) it is best to have a dedicated server socket for each client. Therefore, I am following the answer to this SO question:
Server Listening on Multiple Ports [Java]

Handling Multiple TCP Connections In Java (Server-side)

I'm in the process of writing a messaging program, and I'm running into a spot where I'm having trouble understanding how to pass a socket over to a new thread for handling outbound messages via TCP. I'm currently using UDP packets for messages coming from a client, to the server, which, being UDP, doesn't require very much processing, as it's simply listening for incoming packets, before it de-serializes the objects, and processes them as needed in a separate thread. My problem now is, I'm setting up a client initiated TCP socket for reverse traffic, from the server to the assorted clients that connect. I've done a bit of research, and I already understood that each client should have their own thread for handling outgoing messages, along with another thread simply for accepting the incoming connections. I'm unsure of how to actually achieve this, and I've done some research into the topic.
I've found this: http://docs.oracle.com/javase/tutorial/networking/sockets/clientServer.html
The resource above basically verified my original suspicion that this would have to be handled by threads dedicated to the client. They included psuedo code here, which is representing my listener thread.
while (true) {
accept a connection;
create a thread to deal with the client;
}
I'm a bit of a visual learner, and I have been searching for some type of an example where this is done. I'm unsure of what variable I'd be passing over to the thread that keeps the original connection open, and pushes data back to clients. I'm also having a little bit of trouble grasping whether it even keeps the same socket open, or if a new one needs to be established, which then, makes me believe a firewall could interfere, but I know that won't be the case.
Can somebody explain this for me in detail? If possible, an example would be greatly appreciated!
I'll be likely replying and commenting on responses in about 15-30 minutes from the time this is posted.
What you are doing sounds correct. I typically implement a server like this (simplified version with no tracking of the clients and so on):
#Override
public void run() {
//start listening on the port
try {
serverSocket = new ServerSocket(port);
logger.info("Listening for connections on port " + port);
} catch (IOException e) {
logger.error("Cannot start SocketListener on port " + port + ". Stopping.", e);
return;
}
while (!stopped) {
try {
//wait for connection
Socket newSocket = serverSocket.accept();
ClientThread client = new ClientThread(newSocket);
Thread clientThread = new Thread(client, MEANINGFUL_THREAD_ID);
clientThread.start();
} catch ...
}
}
where serverSocket is a ServerSocket instance variable and stopped is a flag I use to stop the listener thread.
So to answer your questions in the comment, you normally pass the Socket object to each client thread so that that thread can work with the input and output stream and handle closing of the socket and so on. Once you "accept" a socket connection, you do not need to recreate the ServerSocket, you simply call .accept() again to start waiting for a new connection.
In most cases, you will need to keep track of all client threads in your server so that you can stop the server gracefully or do broadcasts for example.

Java TCP server socket accepting multiple clients on different ports

I am trying to build a socket that is capable of having multiple TCP connections, at different ports with multiple clients simultaneously.
The situation is my socket has to receive commands from a server and redirect the commands to wireless sensors (one command to one sensor, not broadcasting), and be able to collect the corresponding data from the sensors and then send them back to the server. Connections with the server would be using TCP, but connections with the sensors could be TCP or UDP.
My main concern are the TCP connections. I am looking into java multithreaded or thread pooled socket tutorials. But the examples I’ve seen were only using a single port to handle all the connections. I’m also trying to look into other possible solutions: utilizing tomcat server, java servlet/JSP, socket channel etc…
I’m not an expert in networking or socket programming so I really hope someone with experience could point me in the right direction. Thank you for any help you can provide in this situation.
Not sure if I fully understand but it seems like it is unnecessary for me to obtain multiple ports for my situation. Seems like I would need to focus on multi threaded sockets and Java NIO topics?
And again thank you for the advice and help.
This is rather a big project to be answered in full here. But here are some general guides:
1) If you want to create a socket on one port you need to create one thread to run it. That is called a server socket. Therefore, from the main thread u need to call one thread for every socket on every port.
2) Each server socket keeps listening on a certain port and waits for clients to connect.
when a client actually connects, the server socket should open another thread for that connection alone and return back to listening.
while(myServerSocket.accept())
{
Open connection thread
}
My advice would be to learn online about how to open threads from classes and then follow the guide above.
Unless you are going over 10k connections then most web servers would be able to handle the traffic.
But maybe you should get more details on the difference between a connection, a socket, and a port. Take a look at this: What is the difference between a port and a socket?
On your question: one port can handle many connections. You don't need different ports for different connections.
My understanding of this request is that you would need to deploy several instances of your server socket application listening on their respective ports and capable of servicing multiple client connections.
By way of mathematical induction, if you have written your server app properly it should work anywhere it is deployed. Below is sample of what your server socket application run should look like
public void run()
{
try
{
while(true)
{
try
{
Thread client_thread= new Thread(new ClientReqProcessor(serverSocket.accept()));
client_thread.start();
}
catch(Exception ex)
{
logger.error("connection error: "+ex.getMessage());
}
}
}
}
Why do you want to use multiple ports? You can have multiple connections on one listening port. The connection itself runs always on different ports.
while (running)
{
try
{
#SuppressWarnings("resource")
Socket socket = serverSocket.accept();
new ServerHandler(socket); // ServerHandler is your class to handle one connection.
}
catch (IOException e)
{
...
}
}
When you really need different ports, you can create a lot of Threads, each with a ServerSocket. Ports are one of the most limited resources on your computer. The running Threads are much cheaper.
for(int i = 1000; i < 1100; i++)
{
final int port = i;
new Thread(new Runnable()
{
#Override
public void run()
{
try
{
ServerSocket serverSocket = new ServerSocket(port);
Socket socket = serverSocket.accept();
new ServerHandler(socket); // ServerHandler is your class to handle one connection.
}
catch(IOException e)
{
...
}
}
}).start();
}

MulticastSocket: Socket operation on non socket

I have some code like this:
InetAddress bind = InetAddress.getByName("192.168.0.1")
MulticastSocket socket = new MulticastSocket(new InetSocketAddress(bind,0));
socket.setInterface(bind);
On windows 7 and windows XP with JDK6u17,I got a SocketException: Socket operation on non socket.
But if I change the line 2 to :
MulticastSocket socket = new MulticastSocket(0);
It's ok, and works find too with jdk6u14.
Why? thanks.
EDIT:
Why port 0 should be the matter?
MulticastSocket socket = new MulticastSocket(0);
Everything goes well with this code.But not
MulticastSocket socket = new MulticastSocket(new InetSocketAddress(bind,port));
Whatever the port is.
As you are binding to a specific interface, calling setInterface() to the same interface is redundant. Remove it. It's only needed when you bind to INADDR_ANY, or in Java an InetAddress of null (or unspecified as a parameter).
To address errors in some of the other answers, and their implications:
Port zero is legal. It means a system-assigned port.
You only need a MulticastSocket for receiving multicasts. For sending, you can just use a DatagramSocket.
If the multicast interface needs to be specified, which it doesn't in this case, it can be done either via MulticastSocket.setInterface() or when calling joinGroup() or leaveGroup(). The latter option gives you granularity at the group level, but both techniques work. That's why they're both provided.
If you don't bind to a specific interface you should definitely call setInterface(). If you are on a multi-homed host you must to call joinGroup()/leaveGroup() once per interface, if you want to receive via all of them.
And a question: is 192.168.0.1 an IP address of an NIC on the local machine? It needs to be.
According to the documentation, you are supposed to instantiate it with a port number (thus 0 would be valid).
I am not so sure.
What's the constructor MulticastSocket(SocketAddress bindaddr) for.
And why it works fine with jdk6u14,but not jdk6u17?
And why it ok on windows 2003 server with jdk6u17?
On RHEL5.2 jdk1.4+
http://www.sockets.com/err_lst1.htm
Berkeley description: An operation was attempted on something that is not a socket. The specified socket parameter refers to a file, not a socket.
WinSock description: Same as Berkeley. The socket input parameter is not a valid socket handle (either it never was valid, it's a file handle (not a socket handle), or if it was a socket handle, it has been closed).
Detailed description:
select(): fails with WSAENOTSOCK if any socket in an fd_set is an invalid socket handle.
Developer suggestions: Did you close a socket inadvertently in one part of an application without keeping another part notified? Use socket state in an application and/or handle this error gracefully as a non-fatal error.
when the MulticastSocket created,socket.isClosed()==true
I haven't used these classes before, but the Exception occurs on line 3 when you call the setInterface method.
I would guess it's something to the effect that you're using the same reference twice or something.
I found a snippet of code that looked like this, maybe this is how you should be doing it:
MulticastSocket ms = new MulticastSocket(new InetSocketAddress(0));
ms.setInterface(InetAddress.getByName("192.168.0.1"));
You should first create the Multicast socket with a well known port - something higher than 1024 and less than 65535 - as already stated 0 means the operating system will choose a port for you (but then its going to be kinda random - which I guess you don't want).
For multicast - you generally need to set the interface to use on joinGroup() not on creation - e.g:
MulticastSocket socket = new MulticastSocket(2121);
InetSocketAddress socketAddress = new InetSocketAddress("localhost", 2121);
if (networkInterfaceName != null){
NetworkInterface ni = NetworkInterface.getByName(networkInterfaceName);
socket.joinGroup(this.socketAddress, ni);
}else {
socket.joinGroup(socketAddress.getAddress());
}
According to the MulticastSocket documentation you should use
Class D IP addresses in the range
224.0.0.0 to 239.255.255.255, inclusive
to bind a MulticastSocket. Apparently, the "192.168.0.1" is out of the multicast range.

Categories