tornado 消费队列并配合 superviosr 优雅重启

superviosr.conf

[group:tornadoes]
programs=ana-8000,ana-8001,ana-8002,ana-8003

[program:ana-8000]
directory = /root/precedents_auto_clean
command = python3 run.py
autorestart = true
restart=true


[program:ana-8001]
directory = /root/precedents_auto_clean1
command = python3 run.py
autorestart = true
restart=true

[program:ana-8002]
directory = /root/precedents_auto_clean2
command = python3 run.py
autorestart = true
restart=true


[program:ana-8003]
directory = /root/precedents_auto_clean3
command = python3 run.py
autorestart = true
restart=true

消费队列

具体参考 pika python rabbitmq 优先级队列

http_client = tornado.httpclient.HTTPClient()
    try:
        time_out = 100
        #port在各个项目中配置好
        url = "http://127.0.0.1:{}/clean?id={}".format(port, id)
        req = httpclient.HTTPRequest(url, request_timeout=time_out)
        http_client.fetch(req)

    except httpclient.HTTPError as e:
        if e.message == "Internal Server Error" or e.message == "Timeout":
            # 杀掉当前的port进程
            command = '''kill -9 $(netstat -nlp | grep :''' + str(
                port) + ''' | awk '{print $7}' | awk -F"/" '{ print $1 }')'''
            os.system(command)
            print("杀死" + e.message)
            # 这里sleep 5s 是为了等待 supervisor autorestart
            time.sleep(5)

    http_client.close()

自启动脚本

参考 screen + rc.local 实现开机自启动多个窗口命令

启动 tornado 的部分

/usr/bin/supervisord -c /etc/supervisord.conf

启动消费队列

screen_name="c1"
screen -dmS $screen_name
cmd1="python3 /root/precedents_auto_clean/cron_receive.py";
screen -x -S $screen_name -p 0 -X stuff "$cmd1"
screen -x -S $screen_name -p 0 -X stuff '\n'