基于RabbitMQ的RPC通信

基于RabbitMQ的RPC通信

当需要调用局域网中的服务时,可以用frp进行穿透,也可以在公网搭建RabbitMQ服务器做消息中转,本文演示了这个步骤。

版本信息

属性 值
RabbitMQ 3.13.0
pika 1.3.1
python 3.6.8
CPU 2颗 Intel® Xeon® Gold 6248R CPU @ 3.00GHz

操作步骤

搭建RabbitMQ(默认用户名:guest 密码:guest )

bash 复制代码
docker run -d --name rabbitmq -p 5671:5671 -p 5672:5672 -p 4369:4369 \
				-p 25672:25672 -p 15671:15671 -p 15672:15672 rabbitmq:management

服务端实现(srv.py)

python 复制代码
# -*- coding:utf-8 -*-
import pika
import time
import sys
import os

def request_handler(message):
    #content=message.decode('utf-8')
    #print(content)
    return message
    
def on_request(ch, method, props, body):
    response = request_handler(body)
    ch.basic_publish(exchange="",
                     routing_key=props.reply_to,
                     properties=pika.BasicProperties(
                         content_encoding='utf-8',
                         correlation_id=props.correlation_id,
                     ),
                     body=str(response))
    ch.basic_ack(delivery_tag=method.delivery_tag)

while True:
    try:
        credentials = pika.PlainCredentials("guest", "guest")
        connection = pika.BlockingConnection(pika.ConnectionParameters("127.0.0.1",port=5672, credentials=credentials))
        channel = connection.channel()
        channel.queue_declare(queue="rpc_queue")    
        channel.basic_consume("rpc_queue", on_request)
        print(" waiting requests")
        channel.start_consuming()
    except:
        print("error restart")
        pass
    time.sleep(1)

客户端实现(client.py)

python 复制代码
# -*- coding:utf-8 -*-
import os
import io
import time
import uuid
import sys
import requests
import warnings
import numpy as np
import pika
import time
import sys
import traceback
import json
import queue
import uuid
import time
import argparse

class MqRpcClient(object):
    def __init__(self):
        credentials = pika.PlainCredentials("guest", "guest")
        self.connection = pika.BlockingConnection(pika.ConnectionParameters("127.0.0.1",port=5672, credentials=credentials))
        self.channel = self.connection.channel()
        result = self.channel.queue_declare(queue="", exclusive=True)
        self.callback_queue = result.method.queue
        self.channel.basic_consume(self.callback_queue, self.on_response, True)

    def on_response(self, ch, method, props, body):
        if self.corr_id == props.correlation_id:
            self.response = body

    def send(self, message,timeout=5):
        self.response = None
        self.corr_id = str(uuid.uuid4())
        self.channel.basic_publish(exchange="",
                                   routing_key="rpc_queue",
                                   properties=pika.BasicProperties(
                                       reply_to=self.callback_queue,
                                       content_encoding='utf-8',
                                       correlation_id=self.corr_id,
                                   ),
                                   body=message.encode('utf-8')) 
        t0=time.time()
        while self.response is None:
            self.connection.process_data_events()
            t1=time.time()
            if t1-t0>timeout:
                return None
        return self.response

def send_message(message,loop):
    rpc = MqRpcClient()
    t0=time.time()
    for i in range(loop):
        response = rpc.send(message,1)
        if response is None and response==message:
            print("Failed")
            break
    t1=time.time()
    print("qps:{:.2f}".format(loop/(t1-t0)))
    return True

if __name__ == "__main__":
    parse = argparse.ArgumentParser()
    parse.add_argument('--msg', type=str, default='', help='')
    parse.add_argument('--loop', type=int, default=1, help='')
    args = parse.parse_args()
    send_message(args.msg,args.loop)

性能测试(4919 qps)

bash 复制代码
# 安装依赖
pip3 install pika==1.3.1

# 启动服务端(服务端收到请求后,直接回复)
python3 srv.py &

# 客户端多路请求,测试总吞吐
cat <<EOF | tee rpc.sh
for i in \`seq 0 \$1\`
do
  python3 client.py  --msg HelloWorld --loop 1000 &
done
wait
EOF
bash rpc.sh  16 | awk -F: '{SUM+=$2};END{print SUM}'
相关推荐
骑着蜗牛撵大象3271 天前
飞牛fnOS部署Damselfly:Compose编排、照片目录与7100固定访问
java·docker·rpc·dubbo·fnos·照片管理·damselfly
syagain_zsx3 天前
BRpc 二次封装:Channel 管理、服务管理与回调工厂
c++·学习·程序人生·rpc·谷歌
mister_guo3 天前
HTTP 和 RPC:从一次干饭说清它们的区别
网络协议·http·rpc
zzu123zsw3 天前
面对多 RPC 扇入超时崩溃:深入 asyncTool 核心 DAG 编排与线程池隔离防护
java·网络·spring boot·网络协议·rpc·并发编程
syagain_zsx3 天前
BRpc 搭建简单 rpc 服务器
服务器·c++·网络协议·rpc·protobuf·库
catino4 天前
RabbitMQ详解
分布式·rabbitmq·ruby
SatanII5 天前
K8s 存储与配置管理|Volume、ConfigMap、Secret 核心笔记
笔记·rpc·kubernetes
叶总没有会5 天前
RabbitMQ高级
分布式·rabbitmq
斌糖雪梨5 天前
registerListeners(); 方法详解
网络·网络协议·rpc
笨鸟先飞的橘猫6 天前
skynet——RPC 超时机制优化方案
游戏·rpc·skynet