Supervisor-php fpm导致502坏网关

我有一个基于php和nginx图像的web应用程序。。。在PHP配置下设置命令之前,一切都很正常: command: /usr/bin/supervisord -c /symfony/supervisord.conf

docker-compose.yml公司

version: '2'

services:
    php:
        build: docker/php
        tty: true
        volumes:
            - '.:/symfony'
        command: /usr/bin/supervisord -c /symfony/supervisord.conf

    nginx:
         image: nginx:1.11
         tty: true
         volumes:
             - './public/:/symfony'
             - './docker/nginx/default.conf:/etc/nginx/conf.d/default.conf'
         ports:
             - '80:80'
         links:
             - php

server {
 server_name ~.*;

 location / {
     root /symfony;

     try_files $uri /index.php$is_args$args;
 }

 location ~ ^/index\.php(/|$) {
     client_max_body_size 50m;

     fastcgi_pass php:9000;
     include fastcgi_params;
     fastcgi_param SCRIPT_FILENAME /symfony/public/index.php;
 }

 error_log /var/log/nginx/error.log;
 access_log /var/log/nginx/access.log;
}

这是我的上司

[unix_http_server]
file=/tmp/supervisor.sock

[supervisord]
logfile=/tmp/supervisord.log
pidfile=/var/run/supervisord.pid
nodaemon=true

Nginx日志显示:

1 connect()失败 (111:连接被拒绝)连接到上游时,客户端: 172.23.0.1,服务器:~。

如我们所见,nginx报告了一个502坏网关错误。如果我删除最后一行CMD,一切正常。如果我移开线路通过 docker-compose exec php bash

知道为什么添加这个命令会导致502坏网关吗??