Working on a python framework for home automation, cameras and monitoring.

• busy with cameras
• next step is victron mppt monitoring

• RabbitMQ -- backend-to-frontend (using MQTT) -- backend-to-device (using AMQP or MQTT)

• Grafana server stats -- RabbitMQ integratition (via prometheus) -- Statsd integration (server stats via graphite) -- Devices stats (??)

The docker-compose.yaml and .env files are in the source folder root. Container images are built on base images (uploaded to dockerhub), to make deployment as quick as possible. Containers communicate using docker network names. .env file is used for both build time and runtime variables.

See: https://dev.to/rohansawant/installing-docker-and-docker-compose-on-the-raspberry-pi-in-5-simple-steps-3mgl

sudo apt-get install -y libffi-dev libssl-dev
sudo apt-get install -y python3 python3-pip
sudo apt-get remove python-configparser
sudo pip3 install docker-compose

or, Locally:

sudo apt-get install curl docker.io
sudo curl -L "https://github.com/docker/compose/releases/download/1.23.1/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
sudo chmod +x /usr/local/bin/docker-compose
sudo usermod -a -G docker $USER

IMPORTANT: Next, log out and in again in order for user-group permissions to be applied.

or, Inside container:

curl -L "https://github.com/docker/compose/releases/download/1.24.1/docker-compose-$(uname -s)-$(uname -m)" -o /usr/local/bin/docker-compose
chmod +x /usr/local/bin/docker-compose

Install docker-machine: https://docs.docker.com/machine/install-machine/

base=https://github.com/docker/machine/releases/download/v0.16.0 && curl -L $base/docker-machine-$(uname -s)-$(uname -m) >/tmp/docker-machine && sudo mv /tmp/docker-machine /usr/local/bin/docker-machine && chmod +x /usr/local/bin/docker-machine

• list all containers: docker ps
• exec into container: docker exec -ti CONRAINER_ID bash
• kill all containers: docker kill $(docker ps -q)
• delete all stopped containers: docker rm $(docker ps -a -q)
• delete all images: docker rmi $(docker images -q)

• create VM (ubuntu 18.04 LTS -- Digital Ocean):

  docker-machine create --driver digitalocean --digitalocean-image ubuntu-18-04-x64 --digitalocean-region sfo2 --digitalocean-access-token=1aa8b56892d2e8bfc9c4d0a1c55c09109b6fdc8d9ab2e2c26519289c7ed07624 mysmarthome
  

• activate docker env: eval $(docker-machine env mysmarthome) .

• deactivate docker env: eval $(docker-machine env -u)

• ssh into remote machine: docker-machine ssh mysmarthome

• list machines: docker-machine ls

• remove machines: docker-machine rm mysmarthome

• provision a system: docker-machine (create --> ssh --> docker-compose up)

• NOTE: Docker containers may not use volumes/shares/mounts. All shared data must be copied from Dockerfiles

• NOTE: Digital Ocean VMs have to be in the same data center (in this case sfo2) as their floating IPs

• Prometheus: docker-compose exec prometheus /bin/sh
• RabbitMQ: docker-compose exec rabbitmq bash
• Graphite: docker-compose exec graphite sh

Create tables:

flask db init
flask db migrate
flask db upgrade

• currently done at server container start
• look at deployment/webapp/webapp.sh
• python utils/create_users.py --file deployment/webapp/users.json

User file format:

[
    {
        "username": "user1",
        "password": "passw1",
        "role": "administrator",
        "routing_keys": "user1.#"
    },
    {
        "username": "user2",
        "password": "passw2",
        "role": "management",
        "routing_keys": "user2.#"
    }
]

Add user:

from utils.flask_models import User
from utils.flask_app import create_app
from webapp import db

app = create_app()
app.app_context().push()

u = User(username='Arno', email='a.rno')
u.set_password('123456')

db.session.add(u)
db.session.commit()

• RabbitMQ HTTP Auth: https://github.com/rabbitmq/rabbitmq-auth-backend-http/blob/v3.7.x/README.md