Skip to content

Latest commit

 

History

History
189 lines (135 loc) · 6.18 KB

local.md

File metadata and controls

189 lines (135 loc) · 6.18 KB

Local

Dependencies required to run Orga Server

  • Python 2
  • Postgres
sudo apt-get update
sudo apt-get install postgresql postgresql-contrib
  • NodeJS if nvm(Node Version Manager) is not installed: using cURL:
curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.32.1/install.sh | bash

or Wget:

wget -qO- https://raw.githubusercontent.com/creationix/nvm/v0.32.1/install.sh | bash

run nvm after exporting NVM_DIR:

. "$NVM_DIR/nvm.sh"

Node installation, v6.9.1 is LTS, though you can install other versions as well:

nvm install 6.9.1

Steps

Make sure you have the dependencies mentioned above installed before proceeding further.

  • Step 0 - Clone the Open Event Orga Server repository (from the development branch) and cd into the directory.
git clone -b development https://github.com/fossasia/open-event-server.git
cd open-event-server
  • Step 1 - Install python requirements. You need to be present in the root directory of the project.
sudo -H pip install -r requirements.txt

hint: You may need to upgrade your pip version and install following packages if you encounter errors while installing the requirements.

sudo apt-get install python-dev
sudo apt-get install libpq-dev
sudo apt-get install libffi6 libffi-dev
  • Step 2 - Create the database. For that we first open the psql shell. Go the directory where your postgres file is stored.
sudo -u postgres psql
  • When inside psql, create a user for open-event and then using the user create the database.

For ease development, you should create Postgres user with the same username as your OS account. If your OS login account is john, for example, you should create john user in Postgres. By this, you can skip entering password when using database.

CREATE USER john WITH PASSWORD 'start';
CREATE DATABASE oevent WITH OWNER john;
  • Once database is created, exit the psql shell with \q followed by ENTER.

  • Step 3 - Create application environment variables.

cp .env.example .env

The URL is short, thanks to the resemble of Postgres user and OS user.

  • Step 4 - Start the postgres service.
sudo service postgresql restart

for mac users:

brew services restart postgresql
  • Step 5 - Create the tables. For that we will use create_db.py.
python create_db.py
# enter email and password
python manage.py db stamp head

Note: In case you are using Anaconda distribution for python, you may get an import error regarding celery.signals module. Please use the default python version while executing these steps in that case.

  • Step 6 - Start the application along with the needed services.
# Install and run redis
# For Ubuntu, Debian and alike
sudo apt-get install redis-server
# For Fedora, RedHat, CentOS
sudo dnf install redis-server

# Run Celery
# socketio has problems with celery "blocking" tasks
# also socketio is not used in a celery task so no problem to turn it off
INTEGRATE_SOCKETIO=false celery worker -A app.celery

# run app
python manage.py runserver
  • Step 7 - Rejoice. Go to localhost:5000 in your web browser to see the application live.

Flask-SocketIO development

Flask-SocketIO has been used in the project for displaying real-time notifications to the user. Although it's switched off by default. To integrate SocketIO you must set the INTEGRATE_SOCKETIO variable to true at bash.

export INTEGRATE_SOCKETIO="true"

The development server is the one that Flask ships with. It's based on Werkzeug and does not support WebSockets. If you try to run it, you'll get a RunTime error, something like: You need to use the eventlet server. . To test real-time notifications, you must use the Gunicorn web server with eventlet worker class.

If you've installed development requirements, you should have both gunicorn and eventlet installed. To run application on port 5000, execute the following instead of python manage.py runserver:

gunicorn app:app --worker-class eventlet -w 1 --bind 0.0.0.0:5000 --reload

-w specifies the number of worker classes to be used. --reload is used for development environments, so the server is restarted if any of the application python files change.

Now you should be able to access the website at localhost:5000.

Nginx

Gunicorn shouldn't be serving static files, it's supposed to run just the Flask application. You can use Nginx to serve static files and bypass other requests to the Gunicorn server, using it as a reverse proxy server. Proper configuration to enable proxying of WebSocket requests can be found in the Flask-SocketIO documentation: https://flask-socketio.readthedocs.io/en/latest/ (search for Nginx).

For Vagrant Machine

Doing the same for Vagrant machine requires some more configuration. If you're using the Vagrantfile provided in the repo, then you can check that the port forwarding is done as: 8001 -> 5000. So accessing the 8001 port in host machine will access the port 5000 in the guest (vagrant) machine. So in the guest machine, you need to run Nginx at port 5000 and gunicorn at some other port (let's assume port 5001).

map $http_upgrade $connection_upgrade {
    default upgrade;
    ''      close;
}

server {
    listen       5000;

    sendfile off;

    location /static {
        alias /vagrant/app/static;
    autoindex on;
    }

    location / {
    proxy_pass http://127.0.0.1:5001;

    proxy_redirect http://127.0.0.1:5001/ http://127.0.0.1:8001/;

    proxy_http_version 1.1;
    proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection $connection_upgrade;

    break;
    }
}

You can directly use this configuration and put it inside sites-available (/etc/nginx/sites-available/nginx.conf) and create a symlink for it in sites-enabled (/etc/nginx/sites-enabled/nginx.conf).

Test the Nginx configuration and restart the Nginx server. Then run the Gunicorn server.

sudo service nginx testconfig # Should respond with "test is successful"
sudo service nginx restart
gunicorn app:app --worker-class eventlet -w 1 --bind 0.0.0.0:5001 --reload