前言
《Django开发简单Blog系统》系列中,我们已经完成了一个迷你Web项目。那么,怎么把这个项目发布到线上呢?怎样给它一个域名呢?
思路:nginx + uwsgi
环境准备
服务器
阿里云服务器,centos7系统。
python
升级python到3.6.1,统一线上和本地python环境。
1、下载python3.6.1源码
wget https://www.python.org/ftp/python/3.6.1/Python-3.6.1.tar.xz
2、解压源码
xz -d Python-3.6.1.tar.xz
tar -xvf Python-3.6.1.tar
3、编译源码
mkdir /usr/local/python3
cd Python-3.6.1
./configure --prefix=/usr/local/python3 --enable-optimizations
make && make install
如果编译失败,需要先更新编译环境:
gcc -v
g++ -v
yum install gcc
yum install gcc-c++
注:我的环境版本为 gcc version 4.8.5 20150623 (Red Hat 4.8.5-11) (GCC) 。
4、替换python
cd /usr/bin
mv python python.bak
ln -s /usr/local/python3/bin/python3.6 /usr/bin/python
ll python*
python -V
5、解决遗留问题
所有python相关的应用,如果使用/usr/bin/python开头的脚本,替换为/usr/bin/python2.7。比如:
vim /usr/bin/yum
vim /usr/libexec/urlgrabber-ext-down
uwsgi
pip install uwsgi
编写测试:
# test.py
def application(env, start_response):
start_response('200 OK', [('Content-Type','text/html')])
return [b"Hello World"]
启动测试:
uwsgi --http :8001 --wsgi-file test.py
报错:uwsgi: command not found,看来我们需要把python3/bin加入到path。
vim /etc/profile
,在文件最底部找到PATH,添加:
:/usr/local/python3/bin
使配置生效:source /etc/profile
访问 http://ip:8001 ,即可看到Hello World 。
nginx和mysql
参考《在CentOS7上配置PHP运行环境》,安装好了nginx和mysql。
项目部署
代码准备
1、克隆项目到服务器
git clone https://github.com/voidking/djsite.git
2、安装django
pip install django
3、安装pymysql
pip install pymysql
数据库准备
1、创建数据库
# mysql -uroot -p
mysql> create database `djsite` default character set utf8 collate utf8_general_ci;
mysql> exit;
2、修改djsite/djsite/settings.py中的数据库配置
vim djsite/djsite/settings.py
3、创建表结构
python manage.py makemigrations
python manage.py migrate
报错:
django.db.utils.InternalError: (1665, 'Cannot execute statement: impossible to write to binary log since BINLOG_FORMAT = STATEMENT and at least one table uses a storage engine limited to row-based logging. InnoDB is limited to row-logging when transaction isolation level is READ COMMITTED or READ UNCOMMITTED.')
修改mysql的binlog格式为混合模式:
# mysql -uroot -p
mysql> set global binlog_format=mixed;
删除数据库djsite中的所有表,然后再次执行:
python manage.py migrate
启动项目
数据库问题
cd djsite
python manage.py runserver
报错:
File "/usr/local/python3/lib/python3.6/site-packages/django/db/backends/mysql/base.py", line 36, in <module>
raise ImproperlyConfigured("mysqlclient 1.3.3 or newer is required; you have %s" % Database.__version__)
django.core.exceptions.ImproperlyConfigured: mysqlclient 1.3.3 or newer is required; you have 0.7.11.None
解决办法:
vim /usr/local/python3/lib/python3.6/site-packages/django/db/backends/mysql/base.py
进入vim命令模式,输入/version
,按N查找下一个,找到:
if version < (1, 3, 3):
raise ImproperlyConfigured("mysqlclient 1.3.3 or newer is required; you have %s" % Database.__version__)
注释掉它,问题解决。
url问题
cd djsite
python manage.py runserver
再次报错:
File "/root/djsite/djsite/urls.py", line 21, in <module>
url(r'^blog/', include('blog.urls', namespace='blog')),
File "/usr/local/python3/lib/python3.6/site-packages/django/urls/conf.py", line 39, in include
'Specifying a namespace in include() without providing an app_name '
django.core.exceptions.ImproperlyConfigured: Specifying a namespace in include() without providing an app_name is not supported. Set the app_name attribute in the included module, or pass a 2-tuple containing the list of patterns and app_name instead.
解决办法:
vim /usr/local/python3/lib/python3.6/site-packages/django/urls/conf.py
找到:
if namespace and not app_name:
raise ImproperlyConfigured(
'Specifying a namespace in include() without providing an app_name '
'is not supported. Set the app_name attribute in the included '
'module, or pass a 2-tuple containing the list of patterns and '
'app_name instead.',
)
注释掉它,问题解决。
查看效果
cd djsite
python manage.py runserver
启动成功,在服务器上测试访问:
curl localhost:8000/blog/index
使用浏览器查看 http://ip:8000/blog/index ,却无法访问。这是因为在settings.py中,ALLOWED_HOSTS的配置为:
ALLOWED_HOSTS = []
官方文档说:
When DEBUG is True and ALLOWED_HOSTS is empty, the host is validated against ['localhost', '127.0.0.1', '[::1]'].
修改ALLOWED_HOSTS的配置为:
ALLOWED_HOSTS = ['*']
然后启动命令改为:python manage.py runserver 0.0.0.0:8000
,此时即可在浏览器看到部署好的项目。
如果还是不能访问,尝试先关闭防火墙:systemctl stop firewalld
nginx配置
1、首先,在万网上配置域名解析,添加A记录,解析到阿里云服务器IP。假设解析好的域名为django.voidking.com。
2、在nginx的vhost中,添加django.voidking.com.conf,内容为:
server {
listen 80;
server_name django.voidking.com;
charset utf-8;
location /{
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
client_max_body_size 1024m;
client_body_buffer_size 128k;
client_body_temp_path data/client_body_temp;
proxy_connect_timeout 90;
proxy_send_timeout 90;
proxy_read_timeout 90;
proxy_buffer_size 4k;
proxy_buffers 4 32k;
proxy_busy_buffers_size 64k;
proxy_temp_file_write_size 64k;
proxy_temp_path data/proxy_temp;
proxy_pass http://127.0.0.1:8000;
}
}
3、重启nginx,./nginx -s reload
4、测试访问
服务器:curl django.voidking.com/blog/index
本地浏览器:http://django.voidking.com/blog/index
至此,django项目已经部署成功,没有用到uwsgi。如果给django添加守护进程,那么我们的部署就接近完美了。那么,uwsgi又能干什么呢,我们继续研究。
uwsgi
一般启动
1、编写wsgi.py文件
编写django_wsgi.py文件,将其放在与文件manage.py同一个目录下。
#!/usr/bin/env python
# coding: utf-8
import os,django
from django.core.handlers.wsgi import WSGIHandler
os.environ.setdefault("DJANGO_SETTINGS_MODULE", "djsite.settings")
django.setup()
application = WSGIHandler()
2、启动项目
uwsgi --http :8000 --chdir ~/djsite/ --module django_wsgi
3、查看启动结果
lsof -i :8000
,ps aux | grep uwsgi
4、测试访问
http://ip:8000/blog/index
此时,页面是没有样式的,也就是说静态资源加载失败。
5、配置静态资源
uwsgi --http :8000 --chdir ~/djsite/ --module django_wsgi --static-map=/static=static
此时,页面样式就正常了。
高级启动
1、新建uwsgi.ini,与manage.py在同一级目录。
[uwsgi]
http = :8000
chdir = /root/djsite/
wsgi-file = django_wsgi.py
static-map = '/static=static'
2、启动uwsgi
uwsgi uwsgi.ini
3、测试访问
http://ip:8000/blog/index
supervisor
关闭shell后,uwsgi服务就很快关闭了。为了让它后台运行,需要让它变成守护进程。
安装supervisor
pip install supervisor
报错,因为supervisor不支持python3:
Supervisor requires Python 2.4 or later but does not work on any version of Python 3. You are using version 3.6.1 (default, Dec 6 2017, 12:03:59)
[GCC 4.8.5 20150623 (Red Hat 4.8.5-11)]. Please install using a supported version.
Command "python setup.py egg_info" failed with error code 1 in /tmp/pip-build-y9wv4fmm/supervisor/
安装pyenv
为了使用supervisor,我们需要python2.7的环境。而多版本python的管理,推荐使用pyenv。
1、安装pyenv套装
curl -L https://raw.githubusercontent.com/pyenv/pyenv-installer/master/bin/pyenv-installer | bash
内容除了包含 pyenv 以外,还包含如下插件:
- pyenv-doctor
- pyenv-installer
- pyenv-update
- pyenv-virtualenv
- pyenv-which-ext
2、路径添加
vim ~/.bash_profile
,添加:
export PATH="/root/.pyenv/bin:$PATH"
eval "$(pyenv init -)"
eval "$(pyenv virtualenv-init -)"
3、使配置立即生效
source ~/.bash_profile
4、查看安装情况
pyenv -v
5、常用命令
- 查看可安装的python版本列表:
pyenv install -l
- 安装指定版本的python:
pyenv install 2.7.13
- 查看已安装的python:
pyenv versions
- 查看当前设为默认的python版本:
pyenv version
安装python2.7环境
1、配置pyenv下载源为本地目录(可选操作,不做的话下载速度会很慢)
mkdir /root/python/ && cd /root/python/
# 设置变量
export PYTHON_BUILD_CACHE_PATH=/root/python
# 设置变量
export PYTHON_BUILD_MIRROR_URL=/root/python
# 查看变量设置
env | grep PYTHON_BUILD_MIRROR_URL
wget https://www.python.org/ftp/python/2.7.13/Python-2.7.13.tar.xz
2、安装python2.7.13
pyenv install 2.7.13
3、改变全局版本
pyenv global 2.7.13
,python -V
附:改变回原版本
pyenv global system
,python -V
4、刷新数据库
python rehash
安装虚拟环境
1、新建supervisor虚拟环境
pyenv virtualenv 2.7.13 supervisor
2、激活虚拟环境
source /root/.pyenv/versions/2.7.13/envs/supervisor/bin/activate supervisor
或者source activate supervisor
3、安装supervisor
yum install supervisor
pip install supervisor
4、生成配置文件
mkdir -p /etc/supervisor/
echo_supervisord_conf > /etc/supervisord.conf
5、修改配置文件
vim /etc/supervisord.conf
,添加:
[include]
files = /etc/supervisor/*.conf
6、运行
/root/.pyenv/versions/2.7.13/envs/supervisor/bin/supervisord -c /etc/supervisord.conf
7、编辑supervisord.service
vi /usr/lib/systemd/system/supervisord.service
,修改为:
[Unit]
Description=Process Monitoring and Control Daemon
After=rc-local.service nss-user-lookup.target
[Service]
Type=forking
ExecStart=/root/.pyenv/versions/2.7.13/envs/supervisor/bin/supervisord -c /etc/supervisord.conf
ExecReload=/root/.pyenv/versions/2.7.13/envs/supervisor/bin/supervisorctl reload
ExecStop=/root/.pyenv/versions/2.7.13/envs/supervisor/bin/supervisorctl shutdown
[Install]
WantedBy=multi-user.target
8、重启supervisor
ps aux | grep supervisord
systemctl stop supervisord
systemctl start supervisord
9、开机启动
systemctl enable supervisord
守护uwsgi
1、在/etc/supervisor中新建djsite.conf文件:
[program:djsite]
command=/usr/local/python3/bin/uwsgi --http :8000 --chdir /root/djsite/ --module django_wsgi --static-map=/static=static
directory=/root/djsite/
startsecs=0
stopwaitsecs=0
autostart=true
autorestart=true
2、重启supervisor
systemctl stop supervisord
systemctl start supervisord
附:重启djsite命令
supervisorctl -c /etc/supervisord.conf restart djsite
3、测试访问
http://ip:8000/blog/index
页面显示正常,至此守护进程配置成功。
4、退出supervisor环境
source deactivate
,守护进程并没有受到影响。
nginx+uwsgi
以上,我们的djsite项目已经通过uwsgi方式启动起来,并且可以保持后台运行。nginx配置不改变的情况下,我们可以正常访问 http://django.voidking.com/blog/index 。此时,nginx作为反向代理,和uwsgi间通过http交互。
接下来,就配置下nginx和uwsgi通过socket结合的方式。原理:用户发送http请求到nginx,nginx通过socket把请求交给uwsgi,uwsgi拿到django的处理结果,通过socket返还给nginx,nginx通过http返回结果给用户。
1、因为nginx和uwsgi通过socket方式交互,我们需要修改uwsgi.ini的配置为:
[uwsgi]
socket = :8000
chdir = /root/djsite/
wsgi-file = django_wsgi.py
static-map = '/static=static'
master = true
processes = 2
enable-threads = true
# daemonize = /root/djsite/uwsgi.log
2、/etc/supervisor/djsite.conf
,修改为
[program:djsite]command=/usr/local/python3/bin/uwsgi uwsgi.ini
directory=/root/djsite/
startsecs=0
stopwaitsecs=0
autostart=true
autorestart=true
3、重启supervisor
systemctl stop supervisord
systemctl start supervisord
4、修改nginx配置djsite.voidking.com.conf:
server {
listen 80;
server_name djsite.voidking.com;
charset utf-8;
location / {
uwsgi_pass 127.0.0.1:8000;
include uwsgi_params;
}
location /static {
alias /root/djsite/static;
}
}
5、重启nginx
./nginx -s reload
6、测试访问
此时,访问 http://ip:8000/blog/index 失败,访问 http://django.voidking.com/blog/index 正常。因为8000端口不再提供http服务,而是一个和nginx连接的socket。
7、static
请问,此时的静态资源,是通过uwsgi获取的?还是通过nginx直接获取的?做一个测试即可,修改uwsgi为:
[uwsgi]
socket = :8000
chdir = /root/djsite/
wsgi-file = django_wsgi.py
# static-map = '/static=static'
master = true
processes = 2
enable-threads = true
# daemonize = /root/djsite/uwsgi.log
此时,uwsgi不再提供静态资源。重启supervisor,页面样式正常,可见,静态资源是通过nginx获取的。之所以可以获取到,是因为我们之前在djsite/settings.py中配置了:
STATICFILES_DIRS = (
os.path.join(BASE_DIR, "static"),
)
小结
至此,django部署完毕,我们实现了三种部署方法:
- nginx + django(http方式)
- nginx + uwsgi(http方式)
- nginx + uwsgi(socket方式)
在此过程中,解决了一些奇怪的bug,学习了升级python的方法,学习了使用pyenv安装多版本python的方法(类似的还有anaconda),学习了给django或者uwsgi添加守护进程的方法,收获颇丰。