I've put this repository together with the intention that it can be used as a teach yourself tutorial for deploying Django with Ansible. Have fun deploying and please feel free to contribute if you think it can be improved.
The workflow is as follows:
- Spin up an EC2 Ubuntu server via AWS or Vagrant - a sample Vagrant file is included
- Provison the server using Ansible
- Choose either RDS from Amazon, PostgreSQL or MYSQL database
- Deploy your Django project from Git and serve it with this stack:
- Optionally install nodejs, mongodb or celery with reddis
This stack comes with useful logging for gunicorn, supervisord, and nginx. It uses logrotate for managing logs and aws-snapshot-tool for rudimentary (full server image) backups. You can also use this playbook to deploy multiple apps to the same server.
*** Just a note - this is still under production and could still benefit from some security improvements and other tweaks. It's highly advised you use your AWS security settings to block all unneeded ports ***
##2 minute quick start
Start by opening the playbook.yml file and providing values for all the non-commented variables listed under vars. The application will be deployed and owned by application_user. You'll need to provide their password hash by running the following command:
python -c "from passlib.hash import sha512_crypt; print sha512_crypt.encrypt('your-password')"
Later if you log into the server, you can easily change to this user with su - [application_user]. When you do this you'll notice the venv is auto-activated and you are automatically in your web app directory.
Finally, open the hosts file and append your server IP underneath [webservers].
Now, run the playbook against your server with the following command:
ansible-playbook playbook.yml -i hosts --private-key=/Path/to/AWS/key/your.pem
Should any task fail, you can make corrections and run again from that task:
ansible-playbook playbook.yml -i hosts --start-at-task='my task name'
#Assumptions
This project assumes you have a live_settings.py in additon to a settings.py and that your application_name variable is named the same as your Django project. You'll also need to edit your wsgi.py file to point to live_settings.py rather than the default settings.py.
It's also assumed your project is set up as below. Other variations are of course possible but you'll have to change the virtualenv_path, git_root, and django_dir variables accordingly.
./media
./static
./requirements.txt
./application_name
./app1
./app2
manage.py
./application_name
settings.py
live_settings.py
wsgi.py
...
Finally, all sensitive settings are enviroment variables which are exported via the postactivate script (in roles/web/templates). The example below, from 2 scoops of Django is a nice example of how you can configure your local/testing/live settings files.
def get_env_variable(var_name):
""" Get the environment variable or return exception """
try:
return os.environ[var_name]
except KeyError:
error_msg = "Set the %s environment variable" % var_name
raise ImproperlyConfigured(error_msg)
DATABASES = {
'default': {
'ENGINE': 'django.db.backends.postgresql_psycopg2',
'NAME': get_env_variable('DB_NAME'),
'USER': get_env_variable('DB_USER'),
'PASSWORD': get_env_variable('DB_PASSWORD'),
'HOST': get_env_variable('DB_HOST'), # Empty for localhost through domain sockets or '127.0.0.1' for localhost through TCP.
'PORT': '', # Set to empty string for default.
},
SECRET_KEY = get_env_variable('DJANGO_SECRET_KEY')
##SSH agent forwarding and cloning your Git repository
You might have some issues with Ansible timing out or giving a public key denied errors when trying to connect to your private repository. This guide is pretty useful for troubleshooting these errors.
##Variable management and Ansible Vault
Ansible Vault allows you to easily encrypt .yml files which you'll want to commit version control. The vars directory is a good place to organise your encrypted variable files.
ansible-vault encrypt foo.yml bar.yml baz.yml #to encrypt existing files
ansible-vault create foo.yml #to create a new encrypted file
ansible-vault edit foo.yml
To run a playbook with encrypted files:
ansible-playbook site.yml --ask-vault-pass
##Ongoing deployments
Use tags to run a bespoke collection of tasks on your hosts
ansible-playbook build/playbook.yml -i build/inventory/hosts --tags='deploy'
Some useful server commands:
sudo supervisorctl restart application_name
Restart nginx, postgres, etc
sudo service restart nginx
##AWS credentials
If you want to use Vagrant to spin up your AWS servers you'll need to get the following details from your AWS account.
- ACCESS_KEY_ID
- SECRET_ACCESS_KEY
Next, if you haven't already, you'll need to generate a keypair and configure the security groups and subnets for your VPC. Having done all this, you should go to the file called 'local_envs' in this directory and fill in all your information. The database variables in this file are intended for your local database installation.
Finally, install Vagrant and this AWS provider for Vagrant
At this stage you should be able to create an EC2 instance via Vagrant.
To verify, source your variables in the 'envs' file and run the following:
vagrant up --provider=aws
vagrant ssh to take a look around, then exit to leave
Back on your host you can checkout your server details with:
vagrant ssh-config
And finally clean up with:
vagrant destroy
Should Vagrant fail to complete your provisioning, you can always just run Ansible picking up from where you left off.
##Vagrant Once you've configured your enviroment and Ansible variables and you've followed the step above you'll be able to both spin up and provision by simply typing:
vagrant up --provider=aws
Thereafter, you'll be able to make ongoing deployments and updates using Ansible.
##To do Dynamic inventories, Development, Staging, Live setups, Better lockdown etc
##Credits I've mainly borrowed from or been influenced by this repo which was inspired by this excellent blogpost.