Django Deploy

For deploy Django project on Linux server.

Features:

1. Automatic

2. Zero downtime

3. Multiple version

4. Easy rollback

Requirements on server

• Nginx
• Anaconda
• Nodejs
• PM2

Quick start

1. Add "deploy" to your INSTALLED_APPS setting like this:

   INSTALLED_APPS = [
       ...
       'deploy',
   ]
   

2. Configure ssh:

   add host config in ~/.ssh/config

   Host host1
       Hostname xxx.xxx.xxx.xxx
       Port 2222
       User xxxx
       ServerAliveInterval 60
       # IdentityFile ~/.ssh/id_rsa_xxxxx
   

   create ssh-key file

   ssh-keygen -t rsa -C "xxxx@xxxx.com"
   

   you could leave a blank for password, when you execute ssh command it will not ask your password again

3. Add definition in settings of Django app

   GIT_URL = 'git@github.com:path/name.git'
   GIT_DEPLOY_BRANCH = 'stable'
   APP_NAME = 'appname'
   
   DEPLOY = {
       "host1": {  # same as ssh-config
           "task_prefix": "app-process",  # prefix of process name
           "home_path": '/path/to/appname/www',  # path of each versions
           "static_path": '/path/to/appname/statics',  # path of statics for each versions
           "conda_path": '/path/to/anaconda3/bin',  # path of anaconda bin 
           "nginx_conf": "/etc/nginx/sites-enabled/appname",  # enabled site config of Nginx
           "fixed_deploy_path": '/path/to/appname/fixed', # use to do migrate
       }
   }
   

4. create PM2 and Nginx configure template

   create directory name of deploy in your base path of Django app

   deploy/
   └── templates
       └── deploy
           ├── ecosystem.config.template
           └── nginx.conf.template
   

   content of ecosystem.config.template

   ...
   

   content of nginx.conf.template