NOTE: The project is not maintained anymore, the code is here for historical / fork purposes.
The goal of the project is to allow average, tech-oriented user to run web applications with ease. The idea is to create an easy to host, technology agnostic private cloud. The ultimate aim is to achieve greater decentralization of web services, such as social networks, file sharing, blog or email.
While many other tools are looking at containers as a way to run massive applications, Puffin concentrates on lightweight ones, each serving just a handful of people.
You can chose to host the applications on Puffin managed platform or on your own server.
Live demo platform is available at puffin.rocks
Puffin consists of two main components - application catalog and interface that provides means to run the applications. Any of them can be used independently - you can run the applications from the catalog directly, and you can use the interface to run your own applications that are not present in the catalog.
To access installed applications from localhost you need to set-up local DNS. There are many alternative solutions to this problem, the simplest one is to add the following lines at the top of your /etc/resolv.conf file:
nameserver 127.0.0.1 options ndots:0
Which can be done by executing the following command as root:
echo -e "nameserver 127.0.0.1\noptions ndots:0\n$(cat /etc/resolv.conf)" > /etc/resolv.conf
Make sure that you disable your other local DNS server, such as dnsmasq, before running Puffin.
Clone git repository
Puffin application catalog is stored as git submodules. When cloning the repo make sure to use --recursive option:
git clone --recursive email@example.com:puffinrocks/puffin.git
Or if you have already cloned the repo then update the submodules in it:
git submodule update --init --recursive
Clone the repository and use Docker Compose:
Puffin server is automatically reloaded on every code change thanks to reload. To rebuild the code after making more substantial change, such as modifying dependencies, run:
Puffin contains several convenience commands to upgrade the database, manage users, execute internal shell, etc. To get a complete list, run:
docker-compose run puffin --help
To deploy Puffin for private needs, for a single user or a limited number of users, use docker-compose-example.yml file as a basis:
cp docker-compose-example.yml docker-compose-production.yml
You need to change SERVER_NAME and VIRTUAL_HOST variables to point to your domain. You also need to set SECRET_KEY variable to a random value.
For a full list of configuration options see puffin/core/config.py.
To send emails from Puffin and the applications you need to configure few environment variables before starting Puffin. It's probably easiest to register to an external email service to avoid being classified as spammer. The variables are (not all are obligatory, see puffin/core/config.py for more details):
MAIL_SERVER MAIL_PORT MAIL_USE_TLS MAIL_USE_SSL MAIL_USERNAME MAIL_PASSWORD MAIL_DEFAULT_SENDER MAIL_SUPPRESS_SEND
On public server you need to configure wildacard DNS record to point to your root domain and all its subdomains.
To instruct Docker to interact with remote server run:
eval "$(docker-machine env [machine-name])"
Finally you can run Puffin:
docker-compose -f docker-compose-production.yml up -d
Initially only "puffin" user with "puffin" password will be created - make sure to change the password before exposing puffin to the outside world. Later you can either allow other users to register themselves on your platform (via SECURITY_REGISTERABLE=True config setting) or create them manually:
docker-compose run puffin user create [login]
(The password will be the same as login, so it should be changed as soon as possible.)
Clustering is currently not supported, but you may run apps on a separate machine than Puffin server itself. To achieve that take a look on MACHINE_* options. You also won't need network sections in your docker-compose file, since the networks will be created automatically on the remote machine.
Application Update & Backup
Application versions are regularly updated. In order to assure than new version doesn't corrupt the data, an automatic backup of all volumes is performed on every application restart.
AGPL, see LICENSE.txt for details.