nginx主配置文件添加配置：

stream {
    log_format proxy '$remote_addr [$time_local] '
                 '$protocol status:$status bytes_sent:$bytes_sent bytes_received:$bytes_received '
                 '$session_time upstream_addr:"$upstream_addr" '
                 '"$upstream_bytes_sent" "$upstream_bytes_received" "$upstream_connect_time"';

    include /etc/nginx/tcp_conf.d/*.conf;
}

虚拟主机的配置

server {
    listen 39000;
    proxy_pass test_backend_for_health;
    proxy_connect_timeout 10s;
    proxy_timeout 10s;
    access_log /var/log/nginx/test.log proxy;
    open_log_file_cache off;
}

upstream test_backend_for_health {
    server 192.168.2.82:10001 max_fails=2 fail_timeout=30s weight=2;
    server 192.168.2.82:10002 max_fails=2 fail_timeout=30s weight=2;
    server 192.168.2.82:10003 max_fails=2 fail_timeout=30s weight=2;
}

应用程序的模仿

python3脚本

# coding: utf8
import socket
import sys
port=int(sys.argv[1])

def simple_tcp_server(host='192.168.2.82', port=port, reply_message="Hello from server!"):
    # 创建一个socket对象
    with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
        s.bind((host, port))
        print(f'Server started on port {port}...')
        s.listen(1)
        while True:
            conn, addr = s.accept()
            with conn:
                print(f'Connected by {addr}')

                while True:
                    data = conn.recv(1024)
                    
                    if not data:
                        break
                    
                    try:
                        print("Received:", data.decode())
                    except Exception as e:
                        print(e)

                    conn.sendall(reply_message.encode())

if __name__ == "__main__":
    simple_tcp_server()

脚本编写好之后，打开3个终端，运行3次python脚本，参数分别为10001/10002/10003。

启动nginx之后，开始模拟做一次tcp请求

telnet localhost 39000

可以看到日志如下：

关闭其中一个python脚本，比如10003这个，多请求几次会发现，nginx的日志变成如下的样子：

 也就是说nginx会先转发一次请求给后端10003，在发现10003失败之后，立即转发给了剩余两台后端应用中的一个，并且把这次请求的过程也给记录了下来，也就是$upstream_bytes_received这个参数打印的内容。