django+uwsgi+nginx+gevent部署

项目工程目录以下：

第一步，测试runserver
cd django3demo1/mysite
python manage.py runserver 0.0.0.0:8002
http://127.0.0.1:8002/polls/ 访问正常

第二步， 配置uwsgi启动
配置文件uwsgi.ini

[uwsgi1]
ini=:base
uwsgi-socket=0.0.0.0:8101
stats=0.0.0.0:8201
pidfile=/mnt/run/django3demo1.pid
;使用supervisor管理进程时不能设置daemonize模式
daemonize = /mnt/logs/uwsgi.log
;logger=file:$(LOG_ROOT)uwsgi1.log
;env=LOG_FILE_SUFFIX=default_01
;代码更改自动链式重启，也可手动touch /mnt/run/reload_lobby1
;touch-chain-reload=/mnt/run/reload_lobby1
enable-threads=true
single-interpreter=true

[base]
;加载指定的python WSGI模块
module=mysite.wsgi
;开启主进程模式
master=true
;设置socket的监听队列大小（默认：100）
listen=100
;设置工做进程的数量，默承认以设置为cpu核数*2
processes=2
;定义日志格式
;logformat=%(addr) - [%(ctime)] %(method) %(uri) %(proto) %(status) %(size) [%(msecs) msecs] [pid:%(pid)] [rss:%(rssM)M] [vsz:%(vszM)M] "%(uagent)"
logformat=%(addr) - [%(ctime)] %(method) %(uri) %(proto) %(status) %(size) %(msecs) msecs "%(referer)" "%(uagent)"
;通过N秒不活跃状态后销毁全部工做进程（这时就进入了cheap模式），只会剩下主进程
;idle=30
;为每一个工做进程设置请求数的上限，超过上限后工做进程就会被回收
max-requests=50000
;顺序重载每一个工做进程，优先加载完的工做进程当即开始工做，须要配合touch-chain-reload参数使用才能生效
lazy-apps=true
;开启内存使用状况报告
memory-report=true
;每一个工做进程的线程数
threads=1
;请求超过harakiri指定的秒数以后，会丢弃这个请求，且当前woker会被回收（即重启woker）
harakiri=45
;当一个请求是“harakiri”杀死的，会记录信息到uwsgi日志里。打开这个选项会输出额外的信息
harakiri-verbose=true

启动uwsgi
uwsgi --ini mysite/uwsgi.ini:uwsgi1
uwsgi --stop /mnt/run/django3demo1.pid
uwsgi --reload /mnt/run/django3demo1.pid

第三步， nginx配置反向代理到uwsgi

upstream mysite_app {
    server 127.0.0.1:7101 weight=5;
}

server {
    listen       8000;
    server_name  _;
    charset utf-8;

    gzip on;
    gzip_min_length 1k;
    gzip_buffers 16 64k;
    gzip_http_version 1.1;
    gzip_comp_level 6;
    gzip_types text/plain application/x-javascript text/css application/xml application/json;
    #gzip_vary on;

    location / {
        include uwsgi_params;
        uwsgi_pass mysite_app;
    }
}

http://127.0.0.1:8000/polls/ 访问正常

第四步， uwsgi启动配置gevent(可选)
uwsgi --gevent 100 --gevent-monkey-patch --ini mysite/uwsgi.ini:uwsgi1

第五步， shell方法启动uwsgi
改为以人类能直观看懂的方式，配置文件bin/h~~~~ttp_server.sh

#!/bin/bash
function usage()
{
cat <<EOF
Usage: bin/http_server.sh {start|reload|stop|gevent-start|restart}

Options:
    [-h|--help]                          : show this help
EOF
}

SCRIPT_PATH=$0
echo $SCRIPT_PATH
WORK_PATH=$(dirname $(dirname $SCRIPT_PATH))
echo $WORK_PATH

while [ "$#" -gt 0 ] 
do
    case "$1" in
    -h|--help)
        usage
        exit 1
        ;;  
    start) 
        count=ps auxf|grep uwsgi.ini|grep -v grep |wc -l`
        if [ $count -gt 0 ]; then
            echo "http server already start ......"
            exit 1
        fi
        uwsgi --ini mysite/uwsgi.ini:uwsgi1
        exit 1
        ;;  
    restart) 
        echo "killing servers ... "
        ps auxf|grep uwsgi.ini|grep -v grep|awk {'print $2'}|sudo xargs kill -9
        echo "wait for http server to exit...."
        sleep 3
        uwsgi --ini mysite/uwsgi.ini:uwsgi1
        exit 1
        ;;  
    gevent-start) 
        uwsgi --ini mysite/uwsgi.ini:uwsgi1
        exit 1
        ;;  
    stop) 
        uwsgi --stop /mnt/run/django3demo1.pid
        exit 1
        ;;  
    reload) 
        uwsgi --reload /mnt/run/django3demo1.pid
        exit 1
        ;;  
    esac
    shift 
done

启动uwsgi
sh bin/http_server.sh start

第六步， 配置supervisor启动uwsgi
配置文件： mysite_supervisor.conf

[program:uwsgi1]
directory = /mnt/hgfs/vmshare/study/python/django3demo1/mysite
environment = PYTHONPATH=./
command = /mnt/venv/python38/bin/uwsgi --ini mysite/supervisor_uwsgi.ini:uwsgi
autostart = true
;startsecs = 30
autorestart = true
startretries = 3
user = zyb
redirect_stderr = True
stdout_logfile = /mnt/logs/supervisor.log
stdout_logfile_maxbytes = 0
stdout_logfile_backups = 0
stdout_capture_maxbytes = 10MB
priority = 31
stopsignal=QUIT

pip3 install supervisorsudo supervisorctl updatesudo supervisorctl restart allsupervisor的具体使用方式就不具体介绍了