This repository contains the complete source code of The Erlangelist site:
If you want to contribute a minor correction to some article (e.g. language and grammar fixes) just make a pull request to the master. If you'd like to make a bigger change, please open a GitHub issue first.
Running the site locally
- Erlang 20
- Elixir 1.6
Optional dependencies are openssl and certbot. It they're missing, you might see some errors, but the site will be working.
- Go to the
sitefolder and fetch
- From the
sitefolder start the site with
iex -S mix phoenix.server
If all went well, the server will listen on the port 20080.
If you have openssl and certbot, the server will self-certify, and also accept https traffic on port 20443.
Deploying to the production server
Note: Don't deploy to a machine where you run some other system, or which is of some other importance to you. Setup script installs some packages and changes the system configuration (e.g. creates a user, modifies iptables rules, ...).
The target host must run Debian 8 and needs to have
sudo installed. It also needs the Internet access, because various packages, docker images, and other dependencies are fetched during the initial setup and subsequent deploys.
There is a support available for starting a local Vagrant box that can act as a "staging" server. You need to first run
cd provisioning/vagrant && VM_IP=192.168.54.54 vagrant up to start the server with a given IP address (feel free to use another IP address).
On your machine you need
ansible, and a
git client. You also need to configure automatic passwordless ssh login to the server machine (under a login that has sudo privilege).
If you started a local Vagrant box, as explained above, you can run
cd provisioning/vagrant && vagrant ssh-config --host ip_of_the_vm to get the ssh configuration for your server.
Filling in the blanks
You need to create
provisioning/git_keys folder that must contain public keys of all allowed deployers.
This step installs required packages and sets up the system. You can run
provisioning/setup-server.sh server_ip network_interface, where network interface is the name of the interface which accepts the requests from the outer world (for example
For example, to set up a local VM with an IP address 192.168.54.54, you can run
provisioning/setup-server.sh 192.168.54.54 eth1.
Setting up the git remote
The deploy is performed by pushing to the git repository on the remote server. During the setup, the
git user is created on the server. You can login as a
git with any of the keys supplied in
provisioning/git_keys folder, as explained above.
In your local git repository, you need to add another remote which will be the deploy target. For example:
git remote add local_deploy firstname.lastname@example.org:erlangelist
If all things are properly configured, you can simply deploy the site via
git push local_deploy master
The first push will run for some time, because it builds all the docker images from scratch. Once everything is done, the site should be listening at the port 80.