Skip to content
This repository has been archived by the owner on Oct 11, 2021. It is now read-only.

mardix/propel-legacy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

****THIS REPO IS NO LONGER BEING MAINTAINED


Propel

About

Propel allows you to deploy multiple Python/PHP/HTML apps (sites) on a single server, run scripts and background workers.

- Why did I create Propel ?

The main reason, was to deploy multiple Flask apps on a single DigitalOcean VM effortlessly. The other reason, was to make it easy to deploy applications.


##TLDR; Install and run Propel (simple Flask example)

  1. On the server you intend to run your applications, install Propel by running the following commands:

      pip install propel
     
      propel-setup
    
  2. CD in the directory that contains your app, and create the following files:

    -app.py

     from flask import Flask
     
     app = Flask(__name__)
     
     @app.route("/")
     def index():
         return "Hello Propel!"
    

    -requirements.txt

     Flask
    

    -propel.yml

     virtualenv:
       name: "mysitename.com"
     
     web:
       -
         name: "mysitename.com"
         application: "app:app"
    
  3. Inside of your app directory run:

     propel -w mysitename.com
    
  4. Now go to http://mysitename.com

  5. Profit! (or your money back)


Features

Deploy Python

Propel allows you deploy multiple Python sites/applications (Flask, Django) by isolating each app into its own virtualenv, using Virtualenvwrapper, then puts it online using Gunicon+Gevent under Supervisor, to make sure it's always up. Then uses NGinx as proxy to access the sites.

- Deploy PHP/HTML

Propel is not limited to only Python app, it can deploy PHP/HTML sites too using PHP-FPM.

- Script and Workers

Besides deploying sites/apps. Propel can run scripts before and after deployment. Run other scripts individually. And background running scripts (workers) with Supervisor.

- Maintenance mode

Propel also has a maintenance mode that will display a maintenance page when the website is down for maintenance or is being deployed.

- Packages

Propel makes use of the following packages:

- Gunicorn

- Supervisor

- Gevent

- Virtualenvwrapper

- PyYaml

- Jinja2

(They will be installed automatically, so no need to do it manually)

Requirements:

- Nginx

- php-fpm

- git (not really required, but good to have)

Install & Setup

pip install propel

After installing propel for the first time, run the following command to setup Supervisor conf and logs directories compatible to propel.

propel-setup

** You may need root admin

You may also need to install some other packages based on your system.

Once done, you should be good to go.


How to use

Propel works in the current working directory (CWD). You must have the file propel.yml in order for it to execute properly.

So you must cd into the directory that contains your propel.yml

Let's say my application is at: /home/myapp.com/www

cd /home/myapp.com/www

From there you can run the commands below:

propel --all-webs

To deploy all websites. It will also run scripts.before_web and scripts.after_web

propel --all-webs

propel -w | --webs [site.com, [site...]]

To deploy site by name. It will also run scripts.before_web and scripts.after_web

propel -w mysite.com mysecondsite.com

propel -s | --scripts [name, [names ...]]

To run a custom script

propel --scripts my_script_name

Run multiple scripts

propel --scripts my_script_name another_script a_third_script

propel -k | --workers [name, [names ...]]

To run workers in the background using Supervisor

propel --workers worker_name another_name

propel -x | --undeploy

To undeploy all. It will remove sites, scripts, workers, and destroy the virtualenv

propel --undeploy

propel -m | --maintenance on|off

(This has been deprecated)

To activate/deactivate the site maintenance page

propel --maintenance on

propel --maintenance off

propel -c | --create [name]

To create a new application bare repository ready to push application to.

propel -c /home/myapp/myapp

# or

# It will place the repository inside of /home/myapp/myapp
propel -c /home/myapp/myapp

propel -r | --reload

To reload Nginx servers and refresh Supervisors config

propel --reload

propel --status

To show the status of all running Propel application. Running applications are active apps on Supervisor.

propel --status

propel --restart

To completely restart all Supervisors processes

propel --restart

propel.yml

propel.yml is a config file that tells propel what to deploy and run.

How does it work ?

WEB

A simple flask app deployment would look something like this:

virtualenv:
  name: "mynewsite.com"

web:
  -
    name: "mynewsite.com"
    application: "run:flask_app"

    nginx:
      aliases:
        "/static": "static"

**For Python application/sites, virtualenv is required. The requirements.txt must also exist to install the necessary packages.

Upon deploying a Python app, Propel will the following:

  • Instantly set the site on maintenance mode. So when a visitor comes, the will a maintenance page

  • create the virtualenv

  • install the requirements.txt in the virtualenv

  • create a random port and assign it to Gunicorn. By default it will assign gevent as the worker-class for Gunicorn, set the numbers of workers and threads.

  • add the Gunicorn command to Supervisor for process monitoring

  • create the Nginx config file for the site and use it as proxy to point to the Gunicorn instance.

  • ...

  • Profit!

A basic PHP/HTML site would have the following structure. PHP-FPM is required.

web:
  -
    name: "my-php-html-site.com"

**For PHP/HTML site we don't need virtualenv.

Upon deploying, it will create the NGINX config file and use PHP-FPM.

Description of the config above

  • virtualenv: A dict of the virtualenv info. It required for Python application

    • name: The name of the virtualenv
  • web: A list of dict website to deploy

    • name: The website name
    • application: The application path in the current directory: 'module:app'. Only for python
    • nginx: A dict of Nginx config
      • aliases: A dict of aliases to use in NGINX. The key is the nginx location, the value is the path
        • "/static": "my-static"

Concerning paths

By default, all paths are relative to the current directory. So let's say you are under: /home/myapp.com/www, setting the aliases: /static: "my-static", it will automatically add the current directory to the path. The alias will turn into: /home/myapp.com/www/my-static

That's how it would be in the Nginx file:

location /static {
    alias /home/myapp.com/www/my-static
}

** To refer an absolute path, you must prefix the path with a slash /

Example of multiple sites

The same above, but with multiple sites and multiple aliases

virtualenv:
  name: "mynewsite.com"

web:
  -
    name: "mynewsite.com"
    application: "run:flask_app"

    nginx:
      aliases:
        "/static": "mytothersite/static"
        "/photos": "mytothersite/photos"

  -
    name: "my-php-html-site.com"
    nginx:
    	root_dir: "html"

For the my-php-html-site.com the root dir is the html directory that's in the current working directory.

The above conf, will deploy multiple sites. The python app will be deployed in the virtualenv, the php/html site will be deployed as is.

WEB: Advanced config

You can add more NGINX and Gunicorn config. Gunicorn config is only for Python app.

virtualenv:
  name: "mynewsite.com"
  rebuild: False
  directory: ""
  pip_options: ""
  
web:
  -
    name: "mynewsite.com"
    application: "run:flask_app"
    remove: False
    exclude: False
    environment: KEY1="value1",KEY2="value2"
    
	# Nginx Config
    nginx:
      server_name: "site1.com admin.site2.com"
      port: 80
      root_dir: ""
      logs_dir: ""
      aliases:
        "/static": "mysite/static"
      force_non_www: True
      force_www: False
      server_directives: ""
      ssl_directives: ""
      ssl_cert: ""
      ssl_key: ""
  
  	# Gunicorn config
    gunicorn:
      workers: 4
      threads: 4
      "max-requests": 100

Virtualenv config

  • name: The name of the virtualenv

  • rebuild: (bool) When True, it will rebuild the virtualenv

  • directory: (path) The virtualenvs directory. By default it is set to /root/.virtualenvs

  • pip_options: (string) String of options to pass to pip. ie: --process-dependency-links --upgrade

Web config

  • name: (string domain) The site's name

  • application: The application to run. It requires virtualenv. Python only.

  • remove: (bool) When True, it will remove the Nginx config file and stop Supervisor if Python, from the server

  • exclude: (bool) When True, it will not try to deploy or re-deploy. It takes precedence over 'remove'.

  • environment: A list of key/value pairs in the form KEY="val",KEY2="val2" that will be placed in the supervisord process’ environment

NGINX config

  • port: 80 by default

  • server_name: (string) Optional if the server name is different than name, or has multiple server_name. Separate multiple server with space

  • root_dir: If provided, it will be used as root dir in nginx, when using PHP/HTML site

  • logs_dir: If provided, it will be used to save the logs

  • aliases: (dict) location/path pair to create aliases

  • force_non_www: (boolean) - If True, it will redirect www to non-www

  • force_www: (boolean) - If True, it will redirect non-www to www

  • server_directives: (multi line text) Extra directives to include in the server block.

  • ssl_directives: (multi line text) Custom ssl directives to include the server block

  • ssl_cert: (path) path to the SSL certificate path.

  • ssl_key: (path) path the the SSL certificate key

Gunicorn config

For Gunicorn, the following values are set by default:

  • workers: (This is calculated based on the total CPU )

  • threads: 4

  • max-requests: 500

  • worker-class: gevent

To disable any of the default values, you can set them as empty or use the desired value

For more config, please refer to: http://docs.gunicorn.org/en/develop/configure.html

Maintenance config

*Deprecated

The maintenance config allows you to set page to show and turn on/off automatically

  • active: (bool) Turn on/off maintenance page

  • page: (path) The maintenance page. If it doesn't exist, it will fallback to the propel default one

  • allowed_ips: (list) List of ips to allow. When allowed_ips is available, it will allow the ips to access the site, but show the maintenance page to all others.


SCRIPTS

Propel allows you to run scripts. Scripts can be run as is, or before and after deployment:

Scripts is a dict of with script name to execute. A script can contain multiple commands

Each command must have the command param, and optionally directory if the script is not being excuted from the same directory. When directory is provided, it will cd into it and run the command

scripts:

  # Run before all
  before_all: 
    - command: "$PYTHON_ENV myscript.py"
    
  # Run after all
  after_all:
    - command: "$PYTHON_ENV myscript.py"
    
  # run before web deployment
  before_web:
    - command: "$PYTHON_ENV myscript.py"

  # run after web deployment
  after_web:
    - command: "$PYTHON_ENV myscript.py"

  # run this script when UNDEPLOYING
  undeploy:
    - command: ""

  # MY_SCRIPT_NAME: Custom script by name
  my_script_name:
    - 
      command: ""
    -
      directory: ""
      command: "$PYTHON_ENV myscript.py"

And to run any of these scripts, you can just do:

propel -s my_script_name

or multiple scripts:

propel -s my_script_name another_script after_all

Scripts command with variables: $PYTHON_ENV and $LOCAL_BIN

As a convenience, there are a few variables to refer to the virtualenv. They allow you to refer to the location without knowing the full path, specially when in virtualenv.

  • $PYTHON_ENV: refers to the Python (ie: /root/.virtualenvs/myvirtualenv/local/bin/python2.7)

  • $LOCAL_BIN: refers to the local bin like (ie: /root/.virtualenvs/myvirtualenv/local/bin/)

  • $CWD: refers to the current working directory

      -
        command: "$PYTHON_ENV manage.py"
    

The command above will execute the manage.py with the virtualenv python.

Config description

  • command: (string) (required) the command to use. You can the $PYTHON_ENV and $LOCAL_BIN variables in it.

  • directory: (string) The directory the command is being executed at. If empty, it will be executed in the current working directory

  • exclude: (bool) When True it will no run or rerun the worker

before_all, after_all, before_web, after_web, before_workers, after_workers

Some pre-made hook to run before and after

before_all: Before setting up everything

after_all: After setting up everything

before_web: Before web deployment

after_web: After web deployment

before_workers: Before workers deployment

after_workers: After workers deployment

scripts:
    before_all:
      - command: "$PYTHON_ENV myscript.py"
    
    after_all:
      - command: "$PYTHON_ENV another-script.py"
        
    before_web:
      - command: "$PYTHON_ENV myscript.py"
    
    after_web:
      - command: "$PYTHON_ENV another-script.py"
   
    before_workers:
      - command: "$PYTHON_ENV myscript.py"
    
    after_workers:
      - command: "$PYTHON_ENV myscript.py"

UNDEPLOY

undeploy will run when undeploying the application. It can be used to clean up etc when removing the site. ie: propel --undeploy

scripts:
  undeploy:
    - command: "time"
    - command: "rm some_path"

Custom Scripts: $SCRIPT_NAME

You can setup your custom scripts to be run manually.

$script_name (with $script_name being the name of the script). It will be run when called manually.

ie: propel -s setup_cron

scripts:
  setup_cron:
    -
      directory: ""
      command: "mysetup_cron_script"

WORKERS

Workers (just like Scripts) are scripts to run continuously in the background via Supervisor.

workers contains a dict of tasks grouped by name, which contain the work to execute.

  # To run this: propel -k managers
  managers:
    -
      name: "manager1"
      command: "$PYTHON_ENV trigger.py
      environment: ""

  # To run this: propel -k jobs
  jobs:
    -
      name: "myworker1"
      command: "$PYTHON_ENV myyworker1.py
      environment: ""
    -
      name: "myworker2"
      directory: ""
      command: "$PYTHON_ENV myyworker2.py"
      environment: ""
      user: ""
      exclude: True  # Prevent this worker from rerunning

Config description

  • name: (string) (required) The name of the worker

  • command: (string) (required) the command to use. You can the $PYTHON_ENV and $LOCAL_BIN variables in it.

  • directory: (string) The directory the command is being executed at. If empty, it will be executed in the current working directory

  • user: (string) The user

  • environment: (string) Environment string

  • exclude: (bool) When True it will no run or rerun the worker. It takes precedence over 'remove'.

  • remove: (bool) When True it will remove the worker from the script


MAINTENANCE

Propel allows you to set your site on Maintenance mode. When visitors come to the site, they will be greeted with a maintenance page to tell them the site is under maintenance.

To manually set the site under maintenance

propel --maintenance on 

// or 

propel -m on

And to remove it

propel --maintenance off

// or 

propel -m off

Propel already has a default page that it will render upon being under maintenance.

Set the maintenance in propel.yml

You can also set the maintenance mode in the propel.yml. This way it will turn on/off maintenance each time you run propel -w

Edit your propel.yml and add the line below.

maintenance:
  active: True 

Set a custom maintenance page

To set your custom maintenance page, add the page under maintenance. The page must be relative to the site's root

maintenance:
  active: True  
  page: "maintenance/index.html" 

So if your site is at: /home/mysite.com, the maintenance page is at: /home/mysite.com/maintenance/index.html

Put site under maintenance, but give full access to certain ips

Sometimes, even if the site is under maintenance, you would like to check everything on it to make sure it works before activate it back again; or you would want to give certain people access before going live.

To do so, Propel allows to set a list of ips you would like to give access while the site is under maintenance

maintenance:
  active: True
  allow_ips: # List of ips to allow to view the site
    - 1.2.4.5
    - 2.3.4.5

Deactivate Maintenance Mode

To deactivate and put the full site back online

maintenance:
  active: False

If the site is under maintenance using propel.yml, propel -m off will not turn off maintenance. You must deactivate it in propel.yml



Some Xtra

propel --git-init $repo_name

To create a git bare repo directory to push content to with git push

cd /home/mydomain.com

propel --git-init www

It will create 3 directories:

www : Where your application content will reside

www.git : A git bare repo

www.logs: A logs directory

So your git path to push directly could be:

ssh://my.ip.address/home/mydomain/www.git

And when you git push it will update the /home/mydomain/www directory

propel --git-push-web $repo_name

It will add the command propel -w in the post-receive hook file so it redeploy the app on each push. Good for Python app.

cd /home/mydomain.com

propel --git-push-web www

propel --git-push-cmd $repo_name [cmd, [cmd...]]

To add custom command to be executed after a git push

cd /home/mydomain.com

propel --git-push-cmd www 'ls -l' 'cd /' ''

Thank you

Mardix :)


License: MIT - Copyright 2015 - 2017 Mardix

About

Deploy multiple Flask/Django (or PHP/HTML) applications on a single server

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages