Common causes
- Apache (apache2 or httpd) is installed and running on port 80
- An old Nginx master process is still running, often from starting nginx manually instead of via systemctl
- Another web server or container (Docker, Caddy, Traefik, a Node app) is published on 80/443
- Duplicate listen directives, such as listen [::]:80 together with ipv6only=off, making Nginx bind the same port twice
- A control panel service or a snap/package that started its own web server
How to fix it
- Find what owns the port. Run sudo ss -ltnp | grep -E ':80 |:443 ' or sudo lsof -i :80. The output shows the process name and PID.
- Stop and disable Apache if it is not needed. Run sudo systemctl stop apache2 && sudo systemctl disable apache2 (httpd on RHEL). If you need both, move Apache to another port such as 8080 and proxy to it.
- Clean up a stray Nginx. If the owner is nginx itself, run sudo systemctl stop nginx; if processes remain, sudo pkill nginx, then start it with sudo systemctl start nginx so systemd tracks it.
- Check Docker and other services. Run docker ps to see containers publishing 0.0.0.0:80. Change their port mapping or stop them.
- Remove duplicate listen lines. Search with grep -R "listen" /etc/nginx/ and make sure only one server block per address uses default_server, and that ipv6only=off is not combined with a separate IPv4 listen.
- Test and start. Run sudo nginx -t, then sudo systemctl start nginx and check systemctl status nginx.
Shell
sudo ss -ltnp | grep -E ':80 |:443 '
sudo lsof -i :80
sudo systemctl stop apache2 && sudo systemctl disable apache2
sudo pkill nginx # only if stray nginx processes remain
sudo nginx -t && sudo systemctl start nginx
systemctl status nginx --no-pagerHow to stop it happening again
- Install only one web server on ports 80/443, or assign clear ports to each
- Start and stop Nginx only through systemctl
- Document which containers publish which host ports