Python UDP聊天应用跨虚拟机通信失败问题求助

我熟悉网络开发，用Python写了个基于UDP的聊天应用，把服务器部署在代理后的新虚拟机时，发现无法接收外部IP的UDP流量。

服务器核心代码（server.py） #

import socket

localIP = ''
localPort = 1993
bufferSize = 1024

UDPServerSocket = socket.socket(family=socket.AF_INET, type=socket.SOCK_DGRAM)
UDPServerSocket.bind((localIP, localPort))
print("UDP server up and listening")

while (True):

  bytesAddressPair = UDPServerSocket.recvfrom(bufferSize)

  message = bytesAddressPair[0]
  address = bytesAddressPair[1]


  clientMsg = "Message from Client:{}".format(message)
  clientIP = "Client IP Address:{}".format(address)
  print(clientMsg)
  print(clientIP)

客户端核心代码（client.py） #

import socket

UDPClientSocket = socket.socket(family=socket.AF_INET, type=socket.SOCK_DGRAM)
socket.setdefaulttimeout(10)
#I know this IP is a placeholder, it's not that way in my actual code
serverAddressPort = ("1.1.1.1", 1993)

bufferSize = 1024

def recvUDP():
  try:
    response = UDPClientSocket.recvfrom(bufferSize)
    return response
  except Exception as e:
    choice = confirm(
      f"Exception '{e}' occured while awaiting server response. Try again?")
    if choice:
      response = recvUDP()


def getUDP(message):
  UDPClientSocket.sendto(str.encode(message), serverAddressPort)
  try:
    response = UDPClientSocket.recvfrom(bufferSize)
    while response is None:
      time.sleep(0.1)
      response = recvUDP()
    return response
  except Exception as e:
    choice = confirm(
      f"Exception '{e}' occured while awaiting server response. Try again?")
    if choice:
      response = getUDP(message)

def connect():

  response = getUDP("initial")

  if response is None:
    sys.exit()

  print("Initial packet recieved")
  
  
connect()

异常现象 #

• 两台机器间无法通过UDP通信
• 服务器用nc -L -p 1993、客户端用curl测试时可正常接收数据包
• 尝试将服务器绑定到公网IP时，socket模块提示无法绑定

1. 确认代理与虚拟机防火墙配置 #

• 检查代理设备（NAT网关、防火墙等）的UDP转发规则：确保外部IP的1993端口已转发到虚拟机内网IP的1993端口，且明确指定转发UDP协议（部分代理默认只处理TCP）。
• 开放虚拟机内网防火墙的UDP 1993端口：
  • Ubuntu/Debian：执行ufw allow 1993/udp
  • CentOS/RHEL：执行firewall-cmd --add-port=1993/udp --permanent && firewall-cmd --reload

2. 修正服务器代码逻辑 #

• 绑定''（即0.0.0.0）是正确操作，代表监听所有网卡，无需绑定公网IP（虚拟机的公网IP实际在代理设备上，本机仅持有内网IP）。可添加代码验证监听状态：

  # 在bind后添加
  print(f"Listening on {UDPServerSocket.getsockname()}")
  

• 原服务器仅接收数据不回复，导致客户端超时等待，需添加响应逻辑：

  # 在print(clientIP)后添加
  msgFromServer = "Message received"
  bytesToSend = str.encode(msgFromServer)
  UDPServerSocket.sendto(bytesToSend, address)
  

3. 修复客户端代码缺陷 #

• 补充缺失的模块导入：

  import time
  import sys
  # 需根据实际使用的confirm函数补充导入，例如tkinter的消息框：
  from tkinter.messagebox import askyesno as confirm
  

• 完善异常分支的返回逻辑，避免无返回值导致的逻辑中断：

  def recvUDP():
    try:
      response = UDPClientSocket.recvfrom(bufferSize)
      return response
    except Exception as e:
      choice = confirm(f"Exception '{e}' occured while awaiting server response. Try again?")
      if choice:
        return recvUDP()
      else:
        return None
  

4. 网络抓包排查 #

• 在虚拟机内执行tcpdump -i any udp port 1993，确认是否收到客户端的UDP数据包：
  • 抓不到包：代理转发规则存在问题
  • 能抓到但服务器无输出：Python代码监听逻辑异常
• 客户端侧也可通过tcpdump验证数据包是否正常发出，未被本地防火墙拦截

内容的提问来源于stack exchange，提问作者MrToenails