Fix "nginx: configuration file test failed"

nginx: [emerg] unexpected "}" in /etc/nginx/sites-enabled/example.com:42
nginx: configuration file /etc/nginx/nginx.conf test failed

Common causes

• A missing semicolon or unbalanced brace, which often shows up as an error on a later line
• A misspelled directive or one from a module that is not compiled in or loaded
• A directive in the wrong context, such as server outside http or location outside server
• A referenced file that does not exist: certificate, key, include or snippet
• Duplicate server_name/listen defaults, duplicate location blocks or a second default_server
• A stray backup file in sites-enabled or conf.d being included alongside the real config

How to fix it

1. Run the test and read the [emerg] line. Run sudo nginx -t. The [emerg] line shows the exact file and line number; open it with sudo nano +42 /etc/nginx/sites-enabled/example.com.
2. Look at the lines just before. An unexpected } or end of file usually means a missing semicolon or brace a few lines earlier. Check that every directive ends with ; and every { has a matching }.
3. Fix unknown directives. Check spelling, then check whether the module is available with nginx -V 2>&1 | tr ' ' '\n' | grep module. Dynamic modules need a load_module line at the top of nginx.conf.
4. Check referenced files. For cannot load certificate or open() failed errors, verify the path with sudo ls -l. If Certbot never issued the certificate, comment out the ssl lines, reload and issue it first.
5. Clean up included files. Remove backups such as default.bak or site.conf~ from sites-enabled and conf.d, since include globs load them too. Use sudo nginx -T to print the full merged config.
6. Reload only after success. When the test passes, run sudo systemctl reload nginx. A failed reload keeps the old config running, but a restart with a broken config takes the site down.

Shell

sudo nginx -t
sudo nginx -T | less          # full merged config with file markers
sudo nginx -t && sudo systemctl reload nginx

How to stop it happening again

• Always chain nginx -t && systemctl reload nginx instead of restarting blindly
• Keep sites-enabled limited to symlinks of real configs, with no backup files
• Store configs in Git so you can diff and roll back a bad change
• Validate edits before copying them to the server