The weird reverse-proxy behavior of the Admin dashboard on NGINX

Hi,

I have my Poste.io image set up, and it works, though the Administration page is acting really, really weird.

For background: - I am running my email server at mail.mywebsitehere.com (not revealing real website in case of security issue) - I am doing a sort of reverse-proxy, because I want mywebsitehere.com to be a WordPress site. - So what I did was I set up Poste.io to run at mail.mywebsitehere.com, though to route all stuff from port 80 to port 82. Then, NGINX takes port 82 and routes it to the mail subdomain. - All of this works, except for the Administration page. I can go to mail.mywebsitehere.com, log in, see my mail, etc. BUT if I click on the "Administration" link, it redirects to 127.0.0.1/admin/login. I can't figure out what is going on. - I can access the admin panel, but I need to type in my server's IP address, then port 82.

Docker-compose.yml

version: '3'

services:
  mailserver:
    image: analogic/poste.io:latest
    container_name: mailserver
    restart: unless-stopped
    hostname: mail.mywebsite.com
    ports:
      - "25:25"
      - "110:110"
      - "82:80"
      - "444:443"
      - "143:143"
      - "587:587"
      - "993:993"
      - "995:995"
      - "4190:4190"
    environment:
      - VIRTUAL_HOST=mail.mywebsite.com
      - HOSTNAME=mail.mywebsite.com
      - HTTPS=OFF
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - /data/nginx/html/.well-known:/opt/www/.well-known
      - /mail/data:/data

Some data from my server.ini

server_hostname = mail.srandm.com
bounce_email = postmaster@mail.srandm.com
alerts_email = admin@srandm.com
default_domain=

My NGINX config:

server {
        listen 80;
        listen [::]:80;
        root /var/www/html;

        # Add index.php to the list if you are using PHP
        # index index.html index.htm index.nginx-debian.html;

        server_name mail.mywebsite.com;

        location / {
                proxy_pass http://127.0.0.1:82/;
        }

    listen 443 ssl; # managed by Certbot
    ssl_certificate /etc/letsencrypt/live/srandm.com/fullchain.pem; # managed by Certbot
    ssl_certificate_key /etc/letsencrypt/live/srandm.com/privkey.pem; # managed by Certbot
    include /etc/letsencrypt/options-ssl-nginx.conf; # managed by Certbot
    ssl_dhparam /etc/letsencrypt/ssl-dhparams.pem; # managed by Certbot


    if ($scheme != "https") {
        return 301 https://$host$request_uri;
    } # managed by Certbot

}

Some images below as explanatory:

Note that I have manually entered the Server IP in image 3 instead of the .com address. It might work with the actual .com address, but I have nginx telling the browser to never connect to my website if there is no https connection (HSTS), so I have no idea if it would work here or not.

Critically, email delivery still works. It appears it is just the Dashboard affected with this weirdness...

Comments (3)