When using RabbitMQ for asynchronous message processing, it is always found that after one or more consumer processes are started, they are automatically disconnected after a period of time, such as a few minutes or 10 minutes, resulting in a large number of messages in the queue and can not be processed.

In order to ensure the continuous high-performance operation of consumers, supervisor is used for supervision, and kill will restart automatically.

Its principle:

rabbitmq enables heartbeat detection after receiving the connection.tune-ok signaling from the client,

rabbitmq will create two processes for heartbeat detection for each tcp connection:

A process that regularly detects whether data is sent on the tcp connection (sending here means rabbitmq sends data to the client). If no data is sent to the client for a period of time, it immediately sends a heartbeat packet to the client (that is, the consumer), and then performs the next detection in a loop;

Another process regularly detects whether there is data reception on the tcp connection. If no data is received within a period of time, it is determined that the heartbeat has timed out, and finally the tcp connection will be closed.

Final solution:

Install and configure the supervisor to guard the consumer process. Once the process is closed, it will restart automatically.   

Environment: Centos

Step 1: Download

1. yum install -y epel-release

2. yum install -y supervisor

Step 2: Configure

cd /etc/
mkdir supervisord.d
echo_supervisord_conf > supervisord.conf
vim /etc/supervisord.conf

#Add the following configuration information
[include]
files = /etc/supervisord.d/*.conf

  Step 3: start the service file (write the file location vim    / etc/init.d/supervisord  )

#!/bin/sh
#
# /etc/init.d/supervisord
#
# Supervisor is a client/server system that
# allows its users to monitor and control a
# number of processes on UNIX-like operating
# systems.
#
# chkconfig: - 64 36
# description: Supervisor Server
# processname: supervisord
 
# Source init functions
. /etc/rc.d/init.d/functions
 
prog="supervisord"
 
prog_bin="/usr/bin/supervisord"
PIDFILE="/var/run/$prog.pid"
 
start()
{
       echo -n $"Starting $prog: "
       ###Note that the following line must have - C / etc / supervisor.conf, otherwise the modified configuration file will not take effect at all!
       daemon $prog_bin -c /etc/supervisord.conf --pidfile $PIDFILE
       [ -f $PIDFILE ] && success $"$prog startup" || failure $"$prog startup"
       echo
}
 
stop()
{
       echo -n $"Shutting down $prog: "
       [ -f $PIDFILE ] && killproc $prog || success $"$prog shutdown"
       echo
}
 
case "$1" in
 
 start)
   start
 ;;
 
 stop)
   stop
 ;;
 
 status)
       status $prog
 ;;
 
 restart)
   stop
   start
 ;;
 
 *)
   echo "Usage: $0 {start|stop|restart|status}"
 ;;
 
esac

3, Execute command  

chmod +x /etc/init.d/supervisord
chkconfig --add supervisord
chkconfig supervisord on
service supervisord start

Step 4: open web view

# If you want to view the managed process through the web, add the following code to listen to 9001, the default user user, and the default password 123

[inet_http_server] 
port=9001
username=user     
password=123

Step 5: create a listening task     (created in vim)   / (etc / Supervisor. D / rabbitmq_worker. Conf file)

[program:rabbitmq_worker]
command=curl http://IP/Delay.Consumer/start; command line operation at startup
autostart=true                                     ;Whether to follow supervisor start-up
autorestart=true                                   ;Whether to restart after hanging, and restart after accidental shutdown, such as kill Drop!
startsecs=10                                       ;How long does it last? The operation is considered successful
startretries=3                                     ;Number of startup attempts
stderr_logfile=/tmp/rabbitmq_worker_err.log        ;Location of standard error output
stdout_logfile=/tmp/rabbitmq_worker_out.log        ;Location of standard output  

Step 6: restart the supervisor service

service supervisord restart

Step 7: check the service startup status

[kingfly@kingfly supervisord.d]# lsof -i:9001
COMMAND    PID USER   FD   TYPE   DEVICE SIZE/OFF NODE NAME
superviso 9452 root    4u  IPv4 40107676      0t0  TCP *:etlservicemgr (LISTEN)

Step 8: view in web mode (login is required after opening, default login account: user, default password: 123)

http://Server IP:9001

If you want to modify the login account password, you can modify it in the configuration file.  

vim /etc/supervisord.conf

  Well, here, the solution and actual combat of guarding the consumer process through Supervisor are completed.

Start the actual combat soon ~ ask if you don't understand

If you want to learn more about the various commands of the Supervisor, see the last one

Supervisor usage details