TCP 通过 IP 连接 Python 中的地址和端口

Question

我们正在尝试在 Python 3.8 中建立 TCP 连接，我们希望每台机器同时成为客户端和服务器，以便每台机器都可以发送和接收消息。 使用此代码，我们可以（有时）发送和接收数据，但我们注意到每台机器都可以向IPs列表中的第一个 ip 地址发送消息，但忽略了 rest。

也许有比这更好的方法来建立 TCP 连接，这样我们就可以互相发送数据了？

代码的所有三个部分（服务器、客户端和共享分发）都附加到每台机器正在运行的 python 程序。

开始分发：

此处启动服务器套接字，我们为要连接的每个 IP 地址启动线程。

def start_distributed_sharing(self):
    self.serv = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    self.serv.bind(('0.0.0.0', 44445))
    self.serv.listen(5)
    MAX_CONNECTION = 5
    IPs = ['{ip-address}', '{ip-address}', '{ip-address}', ]

    client_threads = []
    for ip in IPs:
        client_threads.append(threading.Thread(target=self.client, args=(ip,)))

    for i in range(0, len(client_threads)):
        client_threads[i].start()
    print("Clients is running")

    while True:
        conn, addr = self.serv.accept()
        server_thread = threading.Thread(target=self.server, args=(conn, addr,))
        server_thread.start()
        print("New connection to server created!")

服务器：

每台机器启动自己的服务器，并等待客户端连接

def server(self, conn, addr):
    while True:
        data = ''
        try:
            data = conn.recv(4096)
        except Exception:
            print("Server: Lost a connection... Retrying...")
            time.sleep(5)
            break
        if not data: break
        try:
            data = json.loads(data.decode('utf-8'))
            print(data)
        except Exception:
            print("Server: Could not decode message: ", data)
    conn.close()
    print('Server: client disconnected')

客户：

在这里，客户端尝试使用给定的 IP 地址连接到服务器

def client(self, ip):
    # print(ip)
    self.cli = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
    while True:
        connected = False
        while not connected:
            try:
                print("Client: Connecting to, ", ip)
                self.cli.connect((ip, 44445))
                connected = True
            except Exception:
                print('Client: Could not connect to: ', ip, '. Retrying...')
                time.sleep(5)
        while True:
            time.sleep(2)
            try:
                print("Client: Sending a msg to, ", ip)
                self.cli.send(json.dumps({"id": "{PC1}", "height": self.nodes[self.current_leader].node_stats.lastBlockHeight, "latency": self.nodes[self.current_leader].avgLatencyRecords}).encode('utf-8'))
            except Exception:
                print("Client: Could not send more data to, ", ip)
                break

Answer 1

如果我理解正确，您希望每个（机器/程序）只有一台服务器？ 在这种情况下，我认为每个服务器都需要一个唯一的端口。 或者，如果您希望每个客户端都像客户端/服务器一样与主服务器通信，则可以使用客户端的 recv 方法。

示例 1（向服务器发送消息并等待响应）：

def client(self, ip):
  self.cli = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
  while True:
      connected = False
      while not connected:
          try:
              print("Client: Connecting to, ", ip)
              self.cli.connect((ip, 44445))
              connected = True
          except Exception:
              print('Client: Could not connect to: ', ip, '. Retrying...')
              time.sleep(5)
      while True:
          time.sleep(2)
          try:
              print("Client: Sending a msg to, ", ip)
              self.cli.send(json.dumps({"id": "{PC1}", "height": self.nodes[self.current_leader].node_stats.lastBlockHeight, "latency": self.nodes[self.current_leader].avgLatencyRecords}).encode('utf-8'))
          except Exception:
              print("Client: Could not send more data to, ", ip)
              break

          # Waiting for server response
          response  = self.cli.recv(1024)

现在，如果你想要一个服务器消息事件，你可以像这样创建一个消息处理程序（对于示例来说它不是很干净的代码）：

def on_message(self, msg):
    print(msg)

def message_handle(self):
    while True:
        # Waiting for server message
        msg  = self.cli.recv(1024)
        # check is message if valid
        self.on_message(msg)

def client(self, ip):
    # print(ip)
    self.cli = socket.socket(socket.AF_INET, socket.SOCK_STREAM)

    while True:
        connected = False
        while not connected:
            try:
                print("Client: Connecting to, ", ip)
                self.cli.connect((ip, 44445))
                connected = True
            except Exception:
                print('Client: Could not connect to: ', ip, '. Retrying...')
                time.sleep(5)

    # Connection established start message handler
    handler = threading.Thread(target=self.message_handle))
    handler.start()

    while True:
        time.sleep(2)
        try:
            print("Client: Sending a msg to, ", ip)
            self.cli.send(json.dumps({"id": "{PC1}", "height": self.nodes[self.current_leader].node_stats.lastBlockHeight, "latency": self.nodes[self.current_leader].avgLatencyRecords}).encode('utf-8'))
        except Exception:
            print("Client: Could not send more data to, ", ip)
            break

在这个例子中，每台机器上只有一个服务器和客户端。 然后，您必须在服务器级别管理将消息重定向到目标客户端（识别发送消息的客户端和目标客户端以将消息寻址到正确的客户端）。