将flask部署到服务器

声明，本博客写给自己看的，相当于云笔记，亲爱的陌生人请勿尝试！！！

所需环境

• centos7
• gunicorn
• nginx
• supervisor

升级Python到3.7.6

参考Centos7下python2.7升级至3.6

1.从官网上下载安装包

wget https://www.python.org/ftp/python/3.7.6/Python-3.7.6.tgz

2.解压缩

tar -xvf Python-3.7.6.tgz

3.创建安装目录

mkdir /usr/local/python3

4.编译

cd Python-3.7.6
./configure --prefix=/usr/local/python3
make all
make install

可能会缺少库,如zlib,缺少什么安装什么

yum -y install zlib*
# 下面的命令把可能用到的依赖都装了
yum install gcc-c++ gcc make cmake zlib-devel bzip2-devel openssl-devel ncurse-devel libffi-devel -y

5.进入安装的绝对路径，检查是否安装成功

/usr/local/python3/bin/python3.6 -V

6.修改软连接,让Python指向Python3.7.6
更改原来的软连接

 mv /usr/bin/python /usr/bin/python2.7.5

让Python指向新安装的Python

ln -s /usr/local/python3/bin/python3 /usr/bin/python

7.修改yum,让它指向就Python

sudo vim /usr/bin/yum

修改#! /usr/bin/python为#! /usr/bin/python2.7.5

gunicorn

1. 安装gunicorn

   pip安装gunicorn

   pip install gunicorn
   

   其实不需要nginx，gunicorn便可运行简单的flask应用，创建如下文件

   #main.py
   from flask import Flask
   app = Flask(__name__)
   @app.route('/')
   def index():
   	return 'hello world'
   if __name__ == '__main__':
   	app.run()
   

   然后运行gunicorn

   gunicorn -w 4 main:app -b 0.0.0.0:8000
   

   可能会出现gunicorn command not found, 需要添加软连接
   ln -s /usr/local/python3/bin/gunicorn /usr/bin/gunicorn

   也可能出现[CRITICAL] WORKER TIMEOUT (pid:25867)
   比如我在运行深度学习模型的时候,加载数据会话费大量的时间,因此需要增大gunicorn的超时时间
   gunicorn -w 4 main:app -b 0.0.0.0:8000 -t 600设置超时时间为600秒

   便可在浏览器中看到hello world信息了
   但是gunicorn对静态文件支持不好，所以仍需要使用nginx做反向代理

2. gunicorn关闭操作如下

   pstree -ap|grep gunicorn
   

   得到以下结果

   
     |           |-grep,4698 --color=auto gunicorn
     |   `-gunicorn,4238 /usr/bin/gunicorn -w 4 run_app:app -b 0.0.0.0:8000
     |       |-gunicorn,4243 /usr/bin/gunicorn -w 4 run_app:app -b 0.0.0.0:8000
     |       |-gunicorn,4248 /usr/bin/gunicorn -w 4 run_app:app -b 0.0.0.0:8000
     |       |-gunicorn,4249 /usr/bin/gunicorn -w 4 run_app:app -b 0.0.0.0:8000
     |       `-gunicorn,4250 /usr/bin/gunicorn -w 4 run_app:app -b 0.0.0.0:8000
   
   

   重启gunicorn

   kill -HUP 4238
   

   关闭gunicorn

   kill -9 4238
   

nginx

1. 安装nginx

   sudo yum install nginx -y
   

2. 启动nginx

   service nginx start
   

3. 配置nginx
   向/etc/nginx/nginx.conf添加如下内容

           location / {
           proxy_pass http://127.0.0.1:8000;
           proxy_redirect off;
           proxy_set_header Host $http_host;
           proxy_set_header X-Real-IP $remote_addr;
           proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
           proxy_set_header X-Forwarded-Proto $scheme;
           }
   

4. 重新启动nginx

   service nginx restart
   

5. 再次启动gunicorn便可看到flask应用已经在运行了

supervisor

supervisor能守护进程，当gunicorn挂了之后，能自动重启

1. 安装supervisor

   easy_install supervisor
   mkdir /etc/supervisor
   echo_supervisord_conf > /etc/supervisor/supervisord.conf
   mkdir /etc/supervisor/conf.d
   

   每个应用单独写一个conf，再由supervisord.conf导入
2. 修改supervisor.conf，在末尾加上

   [include]
   files = /etc/supervisor/conf.d/app.conf
   # your_app.conf为你的app配置
   

3. 添加app.conf
   此处的message_app为groram_name

   [program:message_app]
   directory=/home/message
   command=gunicorn -w 4 run_app:app -b 0.0.0.0:8000
   

4. 启动supervisor

   sudo supervisord -c supervisord.conf	#添加配置文件
   sudo supervisorctl start message_app #启动应用
   

5. 常用命令

   sudo supervisorctl reload  #修改配置文件后重新加载
   sudo supervisorctl start app #启动app
   sudo supervisorctl stop app #停止app
   sudo supervisorctl restart app #重启app