supervisor 启动 celery 及启动中的问题

一、前言

本教程重点在于supervisor的配置过程，celery的安装配置请参考其他教程

二、安装supervisor

1.安装命令
pip install supervisor # supervisor目前只支持python2，但是作为容器来说，并不影响监控程序是python3的程序
2.执行 配置文件 生成命令
echo_supervisord_conf > /etc/supervisord.conf
3.创建配置文件需要的文件夹

mkdir /var/log/celery/mkdir /var/log/celery/mkdir /var/log/supervisor/mkdir /var/run/supervisor/

4.更改权限

mkdir /var/log/celery/mkdir /var/log/celery/mkdir /var/log/supervisor/mkdir /var/run/supervisor/

5.编辑配置文件 /etc/supervisord.conf 加入celery配置项（本配置中没使用beat）

; ==================================;  celery worker supervisor example; ==================================[program:celery]; 如果使用的是虚拟环境（virtualenv等）要把celery命令路径写全command=/home/hiveme/env/cznews/bin/celery worker -A cznews --loglevel=INFO; Alternatively,;command=celery --app=your_app.celery:app worker --loglevel=INFO -n worker.%%h; Or run a script;command=celery.shdirectory=/home/hiveme/cznews/  ;项目所在目录user=nobodynumprocs=1stdout_logfile=/var/log/celery/worker.log ;celery输出日志文件stderr_logfile=/var/log/celery/workererr.log ;celery错误输出文件autostart=trueautorestart=truestartsecs=10; Need to wait for currently executing tasks to finish at shutdown.; Increase this if you have very long running tasks.stopwaitsecs = 600; Causes supervisor to send the termination signal (SIGTERM) to the whole process group.stopasgroup=true; Set Celery priority higher than default (999); so, if rabbitmq is supervised, it will start first.priority=1000; ================================;  celery beat supervisor example; ================================;[program:celerybeat]; Set full path to celery program if using virtualenv;command=/home/hiveme/env/cznews/bin/celery beat --schedule /var/lib/celery/beat.db --loglevel=INFO;command=celery beat -A myapp --schedule /var/lib/celery/beat.db --loglevel=INFO ;the ordinary; remove the -A myapp argument if you aren't using an app instance;directory=/home/hiveme/cznews/;user=nobody;numprocs=1;stdout_logfile=/var/log/celery/beat.log;stderr_logfile=/var/log/celery/beaterr.log;autostart=true;autorestart=true;startsecs=10; Causes supervisor to send the termination signal (SIGTERM) to the whole process group.;stopasgroup=true; if rabbitmq is supervised, set its priority higher; so it starts first;priority=999

三、启动supervisor（需要使用root权限）

supervisord -c /etc/supervisord.conf

四、启动过程中的问题（查看supervisord.log和celery的错误日志）

注意重启服务器
1.INFO spawnerr: can't find command 'celery'（supervisord.log）
解决方案：
检查 celery是否安装在虚拟环境中，如果是在配置文件中 需要把所有celery的命令改成 完整的路径

2.INFO spawnerr: unknown error making dispatchers for 'celery': EACCES（supervisord.log）
主要是没有权限操作日志文件
解决方案：
chown youruser:youruser -R /var/log/supervisor/
chown youruser:youruser -R /var/run/supervisor/
chown youruser:youruser -R /var/log/celery/
chown youruser:youruser -R /var/log/celery/

3.exited: celery (exit status 127; not expected)（supervisord.log）
这日志报错，网上是说127是指命令不合法
但是当把命令单独拿到外面执行时也可以执行，进入celery日志查看后发现如下错误：
supervisor: couldn't setuid to 65534: Can't drop privilege as nonroot usersupervisor: child process was not spawned
还是说权限问题，
解决方案：

目前是通过root运行解决的，有更好办法的小伙伴可以和大家一起交流交流

4.supervisorctl: error: <class 'socket.error'>, [Errno 101] Network is unreachable: file: /data/openfalcon/open-falcon/python/lib/python2.7/socket.py line: 575

解决方案

运行supervisorctl时指定配置文件才可以正常运行supervisorctl,否则supervisorctl默认去/etc/supervisord.conf去读取配置，导致错误。

supervisorctl -c your/path/to/your/conf