WIP backend Dockerfiles with OS X support

[yacn]

• note: I removed server/migrations and server/migrations/* from the .gitignore as these will need to be in version control.

Running the Docker Environment

Requirements

• ENVVAR file in root of repo

Process

1. Install Docker
   • OS X: Will also need VirtualBox
2. Install Fig
3. Start Docker daemon
   • Arch:
     • sudo systemctl start docker
     • You can avoid needing to prefix docker/fig commands with sudo
       by adding yourself to the docker group with
       sudo gpasswd -a <user> docker and then logging out and back in.
   • OS X:
     • boot2docker init (only needed once)
     • boot2docker start (starts the VM)
     • $(boot2docker shellinit) (sets necessary environment variables, will
       want to add this to your shell profile/rc file)
   • verify working by running:
     • OS X: boot2docker status
     • [sudo] docker version
     • [sudo] docker run hello-world
     • [sudo] fig --version
4. Build containers
   • run [sudo] fig build
5. Start containers
   • Create a symlink to fig.yml from the corresponding platform-specific
     fig file
     • Arch: ln -s fig.linux.yml fig.yml
     • OS X: ln -s fig.osx.yml fig.yml
   • NOTE: You'll need to rename the folder this repo is in from Neu-Hatch
     to hatch or something, Docker doesn't allow capital letters in container
     names
     • alternatively, you could prefix each fig command with -p hatch,
       but will get annoying.
   • OS X:
     • source ENVVAR
     • run fig up
   • Arch:
     • run [sudo] fig up
     • If you're not part of the docker group, you'll need to prefix the
       commands with sudo. Since some of the environment variables in the
       fig.yml are sourced from the host that fig is running on, you'll
       need to use a small helper script that you can use sudo with that
       will source your ENVVAR file and then run fig up
6. If this was your first time building the DB container, you'll need to run:
   • TODO: Figure out which of these can be moved into the container build step
   • fig run server python manage.py db init
   • fig run server python manage.py db migrate
   • fig run server python manage.py db upgrade
7. The Hatch backend should now be running and available at:
   • Arch: localhost:5000
   • OS X: $(boot2docker ip):5000

Making Changes

Backend (db and server)

• TODO: figure out process involved with data-only container holding postgres
  data.

If you make changes to the back end code, it's easiest to do the following:

1. Run fig stop or press Ctrl+C in the terminal window where fig is
   running to stop the containers.
2. Run fig rm to remove the old containers
3. Run fig build to build new containers with the latest changes
4. Run fig up. You will need to re-run the database migrations (migrate
   and upgrade, init most likely will not need to be run).

Client

This is out dated as the client is moving to a separate repo

If you have the collection of containers running (db, server, client),
and want to test your changes to the client code, you should do the following:

1. Open a new terminal / session in the same directory as this repo
2. Make your changes (they may already have been made)
3. Run fig stop client to stop the existing client container
4. Run fig rm client to remove the existing client container
5. Run fig build client to create a new client container with your most
   recent changes.
6. Run fig up client to start the client container

TODO

• figure out how to avoid publishing ports 5432 and 5000 through container
  linking
• DATABASE PERSISTENCE
  • Will involve some container that simply mounts/provides a volume where
    the database will live.
• SSL certificates
• Audit configuration for security (i.e. no ports exposed unecessarily)
• figure out production deployment
  • In theory should be simple matter of bringing up a server with Docker and
    Fig installed, cloning this repo, and running fig up

[yacn]

Since this is essentially the same as $PWD/server/:/code, may not need to have two separate fig.ymls

[yacn]

refreshing with latest master