First Time Quick Setup
This section describes how to set up the system for the first time. It is assumed you are logged into the server machine that you wish to use, using the user account that you want to run the server with. The entire application has been dockerised, so can in theory be used on any major operating system with minimal modification.
Deploying on fly.io
The Galv team use fly.io to deploy the Galv application. This is a cloud service that allows you to deploy docker containers to a server.
You can find the example configuration file for fly.io in the repository root directory
fly.toml
. This file is used to configure the deployment of the Galv application.
To run the application it’s easiest to use the prebuilt container which accompanies each release on GitHub. The containers are stored in GitHub’s Container Registry.
To launch with fly.io, you will need to:
Install the fly.io CLI tool, which can be found here.
Log in to the fly.io CLI tool using the command
fly login
.Create a new postgres database using the command
fly postgres create
. The name you give it will be needed in step 5.Create a new application using the command
fly apps create
. The name you give it will be needed in step 5.Attach the postgres database to the application using the command
fly postgres attach <postgres_app_name> --app <galv_app_name>
.Set the environment variables by modifying the
fly.toml
file. Most variables have sensible defaults. You will need to ensure that theVIRTUAL_HOST_ROOT
is set to the domain name you wish to use - e.g.galv-app.fly.dev
. You will also need to ensure that theFRONTEND_VIRTUAL_HOST
is set to the domain name you wish to use for the frontend - e.g.galv-frontend.fly.dev
.Set the secret environment variables by running the command
fly secrets set DJANGO_SECRET_KEY=<secret_key>
. Do the same for the superusers’s password withfly secrets set DJANGO_SUPERUSER_PASSWORD=<superuser_password>
.Launch the application using the command
fly deploy
. This will launch the latest version of the Galv backend on the fly.io server by looking for the container image in the GitHub Container Registry. It is possible to build your own version from the source code.
Deploying on EC2 or other cloud services - single service
You can deploy Galv on any cloud service that supports docker containers. The following instructions are for deploying on an AWS EC2 instance.
Launch an EC2 instance with the Amazon Linux 2 AMI.
SSH into the instance.
Install docker.
sudo yum update -y sudo amazon-linux-extras install docker sudo service docker start sudo usermod -a -G docker ec2-user
Create a .env file with the environment variables you need. This can have both the normal and secret environment variables. You can use the .env file in the repository as a template. At the very least, you will need to set:
DJANGO_SECRET_KEY
DJANGO_SUPERUSER_PASSWORD
Postgres connection details using either
POSTGRES_*
variables orDATABASE_URL
VIRTUAL_HOST_ROOT
FRONTEND_VIRTUAL_HOST
Build the container from the GitHub Container Registry.
docker run -d --env-file .env -p 80:80 ghcr.io/galv-team/galv-backend:latest
Deploying from scratch
You need to have docker
, docker-compose
and git
installed and available on your
command-line.
You can find installation instructions for docker
on all major operating systems
here, and for docker-compose
here. For linux hosts, it is useful to be
able to use docker
as a non-root user, and you can find instructions on how to set
this up here. If you don’t,
note that you will need to add sudo ...
in front of every docker
and
docker-compose
command listed below.
Installation instructions for git
for all major OSs can be found
here.
Get the galv source code
First you will need to clone the galv repository using git
:
git clone https://gitlab.com/galv-team/galv-project/galv.git
cd galv
Setup environment variables
The Galv project uses two .env
files, .env
and .env.secret
.
You will already have a .env
file in the repository you cloned, with sensible defaults.
If you’re running a production deployment, you will want to set the value of the
VIRTUAL_HOST_ROOT
to your domain name, e.g. VIRTUAL_HOST_ROOT=example.com
.
This will serve the Galv web application from the root of your domain,
e.g. at http://example.com/
; and the API from the subdomain, e.g. http://api.example.com
.
You will likely also want to enable HTTPS, for which we use LetsEncrypt to generate SSL certificates.
By default, the staging (test) server is used, which generates certificates that are not trusted by browsers.
When your production setup appears to work correctly, you can switch to fetching real certificates
by setting LETSENCRYPT_TEST=false
and restarting the nginx-proxy container.
If you wish to change where the database is saved, you can change the first entry
in .env
, GALV_DATA_PATH
to the directory where you want the postgres database.
Create .env.secret
The second .env
file is a secrets file.
This is not included because you should come up with your own secret values for the
entries within it.
Create the file and edit it so that it has the following keys:
DJANGO_SECRET_KEY
DJANGO_SUPERUSER_PASSWORD
POSTGRES_PASSWORD
All of these values should be unguessable secure passwords.
DJANGO_SECRET_KEY
should be very long and complex, consider 60+ characters
with a mixture of special characters (avoid $ and £), upper- and lower-case letters,
and numbers.
The only one of these you will need to use again will be the superuser password.
If you would like the Django superuser to have a name that is not ‘admin’,
you can also specify DJANGO_SUPERUSER_USERNAME
.
vi .env.secret # could also use nano, emacs, etc.
Build docker images (only when upgrading to a new version of galv)
If you have previously installed and run galv you might already have old docker images already built. To rebuild the images, run the following command:
docker-compose build
Running Galv
You can run the galv server and web application frontend using the following
docker-compose
command from the root folder of the repository.
docker-compose up app
Now view the ‘localhost’ IP address http://127.0.0.1/ in your browser and you should see the Galv login page. This is the web frontend. If you wish to use the frontend from another machine, use the IP address or URL of the server instead.
Creating a user account
It’s not a good idea to do everything with the Django superuser. You can create a new account by using the web interface to POST to the API. Go to the /users/ endpoint and fill in the DRF form.
You’ll need to approve the account at the /activate/ endpoint.
Open up a web browser and go to the mailhog web port (default 8025 in docker-compose.yml
) to see the email,
then copy the activation code.
Go to /activate/?token=[your_code]&username=[your_username]
in your browser to activate the account.