I often have some services running locally, and accessing their web pages via 127.0.0.1: every time feels too clumsy. It happened that a few days ago ICANN reserved .INTERNAL as a top-level domain dedicated to internal networks, so I used this domain for the configuration.

DNS

  • I need to configure the domain *.internal to point to 127.0.0.1. Since the method of hacking /etc/hosts does not support wildcards, I use dnsmasq to run a DNS service locally.
$ sudo pacman -S dnsmasq
$ sudo systemctl enable dnsmasq

+ Add the domain to the configuration

# /etc/dnsmasq.conf

address=/internal/127.0.0.1

+ A small problem encountered when starting the dnsmasq service: 127.0.0.1:53 is occupied. One solution is to run the service elsewhere in `/etc/dnsmasq.conf`. Remember that you need to configure the DNS address in systemd-resolved. Then restart dnsmasq and systemd-resolved. For various possible solutions, see: [link](https://askubuntu.com/questions/191226/dnsmasq-failed-to-create-listening-socket-for-port-53-address-already-in-use)

# /etc/dnsmasq.conf
listen-address=127.0.53.1
bind-interfaces


# /etc/systemd/resolved.conf
[Resolve]
DNS=127.0.53.1

Web Server

  • I need to configure a web server to point different subdomains to the correct ports. nginx is used here.

  • Start the nginx service and add a server block to the configuration.

# /etc/nginx/nginx.conf

http {
	include conf.d/*.conf;
}

# /etc/nginx/conf.d/internal.conf

server {
    listen 80;
    server_name sub1.internal;

    location / {
        proxy_pass http://127.0.0.1:3000;
    }
}

server {
    listen 80;
    server_name *.internal;
    
	return 404;
}

  • Restart the nginx service. Now visiting sub1.internal is redirected to 127.0.0.1:3000.

  • A small problem encountered: the server_name option stopped working, and any subdomain was redirected to the same port. The counterintuitive behavior of nginx server_name