fail2rest has two options that be configured via config.json
* **Fail2banSocket** - The path to the fail2ban socket, can usually be found via `grep socket /etc/fail2ban/fail2ban.conf` you also have to run fail2rest as a user who has permissions to use this socket
* **Addr** - The address that fail2rest is served upon, it is usually best so serve to the loopback, and then allow access via nginx see an example config in the [fail2web](https://github.com/sean-der/fail2web) repository
The default configuration file used by init-scripts is `/etc/fail2rest.json`. You should download the config to your /tmp/ dir and modify it to your needs.
Once you finished editing the configuration file, you should move it from /tmp/config.json to /etc/fail2rest.json executing `mv /tmp/config.json /etc/fail2rest.json`
However, fail2rest is designed to run as a service, so init scripts are provided that allow easy management of fail2rest. They can be found [here](https://github.com/sean-der/fail2rest/tree/master/init-scripts)
Download the appropriate init file your Distribution. You may need to customize your init script to load your config.json, but most scripts default to /etc/fail2rest.json
To run as a service you can either copy or create symlinks for systemd and the fail2rest binary. The systemd file shold be added to /etc/systemd/system/fail2rest.service and the the binary to /usr/bin/fail2rest. This scenario will use the symlinks in order to always use the latest files. You should run the commands with sudo if not logged in as root: