Indicator Registry

Deployments

We currently have a simplified deployment process with limitted scalability using a fabric script (http://www.fabfile.org/). In the future we plan to to use Amazon Elastic Beanstalk (https://aws.amazon.com/es/elasticbeanstalk/) to handle deployment and scalability.

Fabric script

The deployment script is stored on fabfile folder and was based on this project: https://github.com/ashokfernandez/Django-Fabric-AWS

Requirements:

• Install the libraries of fabfile pip install -r fabfile/requirements.txt.
• Private AWS key file to access the instances installed on your SSH folder ~/.ssh/indicatorregistry.pem (check permissions of the file, 400 is expected)
• Private deploy key on your ssh file indicatorregistry_dply and indicatorregistry_dply.pub
• Passphrase of the deployment key

Deployment commands:

A complete description can be found on fabfile/django_fabric_aws.py.

• fab spawn instance
• fab configure_instance
• fab update_packages
• fab deploy
• fab reload_gunicorn
• fab reload_nginx
• fab reload_supervisor
• fab manage:command="management command"

Customizations

In some deployments we need to simplify how we serve the static files or/and the database.

Nginx

A posibility is to use nginx to serve the static folder of the project. We can also use nginx to handle redirects from non www to www domains.

Deployment steps

You can trigger a deploy to the instances listed on fabfile/project_conf.py variable EC2_INSTANCES by executing

fab deploy

this command will update the repository on the instances to the latest version of the branch selected on each one.

Local environment configuration

To install a local environment you will need python 2.7 in virtualenv wrapper

• Create a virtualenv mkvirtualenv indicatorregistry
• Install dependencies pip install -r logicaloutcomes/requirements/common.txt and pip install -r logicaloutcomes/requirements/dev.txt
• Run migrations ./manage.py migrate --settings=logicaloutcomes.settings.dev
• Collect statics ./manage.py collectstatic --settings=logicaloutcomes.settings.dev
• Populate database with indicators ./load_indicators.sh
• Create a superuser ./manage.py createsuperuser --settings=logicaloutcomes.settings.dev
• Run the development server with logicaloutcomes/settings/dev.py as the settings file. A shortcut for this is ./run.sh