nginx: [emerg] bind() to 0.0.0.0:80 failed (98: Address already in use) — how to fix

nginx: [emerg] bind() to 0.0.0.0:80 failed (98: Address already in use)
nginx: [emerg] bind() to [::]:80 failed (98: Address already in use)
nginx: [emerg] still could not bind()

Quick fix: See who has the port: sudo ss -ltnp 'sport = :80'. If it's Apache: sudo systemctl disable --now apache2. If it's an old nginx: sudo systemctl stop nginx; sudo pkill nginx; sudo systemctl start nginx. If it's a container, stop it or change its -p mapping. Use nginx -s reload (or systemctl reload nginx) after config changes instead of starting nginx again.

What it means

nginx tries to open a listening socket for every listen directive. Error 98 (EADDRINUSE) means something already holds that address and port. After a few retries nginx gives up with still could not bind().

Find the process:

sudo ss -ltnp 'sport = :80'
sudo ss -ltnp 'sport = :443'
# alternative
sudo lsof -i :80 -sTCP:LISTEN

Look at the users:(("name",pid=...)) part.

Cause 1: nginx is already running

Very common: nginx was started by systemd, then you ran sudo nginx by hand (or the other way round). The second copy can't bind.

ps aux | grep [n]ginx

To apply config changes, reload the running instance instead of starting another:

sudo nginx -t && sudo systemctl reload nginx

If an orphaned nginx (not managed by systemd) holds the port:

sudo systemctl stop nginx
sudo pkill -x nginx        # or: sudo nginx -s quit
sudo systemctl start nginx

Cause 2: Apache (or another web server) is installed

On Ubuntu/Debian, installing a package can pull in apache2, which starts on port 80.

sudo systemctl disable --now apache2     # RHEL/Fedora: httpd
sudo systemctl restart nginx

Or keep both and run one on a different port (change Listen 80 in /etc/apache2/ports.conf, or listen 8080; in nginx).

Other usual suspects: Caddy, Traefik, lighttpd, HAProxy, Varnish, a control panel, or snap services.

Cause 3: a Docker container publishes the port

docker-proxy in the ss output means a container mapped -p 80:80:

docker ps --format '{{.Names}}\t{{.Ports}}' | grep ':80->'
docker stop <name>

Or change the mapping (e.g. -p 8080:80) and proxy to it from host nginx.

Cause 4: duplicate listen in your own config

nginx can conflict with itself. The classic case is listen [::]:80 ipv6only=off; (a dual-stack socket that also takes IPv4) together with listen 80; in another server block: both want IPv4 port 80, so the second bind fails with error 98. Show the full config and search for listens:

sudo nginx -T 2>/dev/null | grep -n "listen"

Typical clean setup for each server block: listen 80; and listen [::]:80;. Options like default_server, reuseport or ipv6only may appear only once per address:port. Check for leftover files in /etc/nginx/sites-enabled/ and /etc/nginx/conf.d/ too.

Cause 5: port 443 held by something else

Same commands with :443. If a VPN, Skype, or another proxy uses 443, stop it or move it.

After fixing

sudo nginx -t
sudo systemctl restart nginx
systemctl status nginx
curl -I http://localhost

Last updated 2026-11-01.