Sushi bar

The place where all the sushi chefs hang out.

Features

• Dashboard to monitor chefs' progress, logs, and run history
• Remote control of a chefs started in daemon mode

Deploy to prod

Assuming your local repository has latest code + credentials + docker installed:

# setup access to remote docker daemon
eval $(docker-machine env gcpsushibarhost)    # this will set 4 env vars in current shell

docker-compose -f production.yml  build
docker-compose -f production.yml  up -d
docker ps

Settings

• Use config/settings/base.py for common settings
• Use config/settings/local.py for local development settings (default option when running ./manage.py).
• Use config/settings/production.py for added prod security restrictions. Credentials will be sourced from the file .prodenv.

Localhost provision

You'll need to install Python3, Postgres DB, and redis on your machine.

Code:

virtualenv -p python3  venv
source venv/bin/activate
pip install -r requirements/local.txt
./manage.py makemigrations

Create DB for local development (assuming Postgres is running on localhost):

createdb sushibar

Migrate models

./manage.py migrate

For convenience there you can load an predefined admin user fixture using

./manage.py loaddata sushibar/users/fixtures/admin_user.json

Then you can login with username admin and password admin123. Alternatively, you can create a new admin account using:

./manage.py createsuperuser

Run in development

./manage.py runserver

Running tests

./manage.py test runs   # test sushibar api
./manage.py test        # all tests

Clean-slate restart

This will drop all the data in the DB and restart:

dropdb sushibar
createdb sushibar
rm -rf runs/migrations/0*.py
./manage.py makemigrations
./manage.py migrate
./manage.py loaddata sushibar/users/fixtures/admin_user.json

Production setup

Before we set this up as a kubernetes, we can test all the dockerization using the tools docker-machine and docker-compose.

# 1. setup env vars that proxy local docker commands to the docker host `gcpsushibarhost`
eval $(docker-machine env gcpsushibarhost)

# 2. create network
docker network create nginx-proxy

# 3. start all containers
docker-compose -f production.yml up -d

# check what's running
docker ps

# View nginx+django logs (like tail -f)
docker-compose -f production.yml  logs -f nginx django-wsgi

Possibly use https://github.com/kubernetes-incubator/kompose to generate the Kubernetes config from production.yml when it's done.

Update production server

To deploy new code after updating the local repository, run the following steps:

# setup access to remote docker daemon
eval $(docker-machine env gcpsushibarhost)    # this will set 4 env vars in current shell

# rebuild
docker-compose -f production.yml  build

# update running containers
docker-compose -f production.yml  up -d

# check containers are running OK
docker ps

TODO: research what --no-deps flag does and if it's better.

Debugging production setup

See what's going on:

docker ps                                   # running containers
docker ps -a                                # running and stopped containers
docker logs django-asgi                     # see latest logs from the ASGI container
docker-compose -f production.yml  logs      # see latest logs from all containers

Run bash inside container, while allocating a tty and using interactive mode:

docker exec -ti nginx  /bin/bash

See current production nginx config:

docker exec nginx   cat /etc/nginx/conf.d/default.conf

Show all the network info

docker network ls
docker network ls -q | xargs docker network inspect

Show intenral IPs

docker inspect -f '{{.Name}} - {{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' $(docker ps -aq)
    /nginx-gen - 172.18.0.8
    /letsencrypt-nginx-proxy-companion - 172.18.0.9
    /django-wsgi - 172.18.0.7
    /nginx - 172.18.0.6
    /asgi-worker - 172.18.0.5
    /django-asgi - 172.18.0.4
    /sushibar-postgres - 172.18.0.3
    /sushibar_redis_1 - 172.18.0.2

Volumes check

docker volume ls
docker volume rm <volume id>

Restart from scratch

# bring containers down and make sure volumes are deleted
docker-compose -f production.yml  down -v

# cleanup images
docker images
docker-compose -f production.yml  rm
docker rmi -f sample-api sample-website jwilder/docker-gen mhart/alpine-node \
              jrcs/letsencrypt-nginx-proxy-companion nginx sushibar_nginx-gen
docker images

# rebuild
docker-compose -f production.yml  build --no-cache

docker-compose -f production.yml  up  # prints combined strout from all containers