NGINX regexp at the end of location path

Question

I'm trying to set up a location such that suffix /reset_password/any_text would load index.html from the directory specified in the alias. Unfortunately, with no luck...

# location /reset_password { # alias /var/www/grades-ui; # } location /reset_password/([0-9a-z]*) { alias /var/www/grades-ui; }

The commented out code works fine to load exact path /reset_password, however my goal is to serve /reset_password/any_text_or_number.

Any ideas why my regular expression wouldn't work? Or maybe I'm doing something else completely wrong?

Edit: the complete conf file

server { listen 443 ssl; # managed by Certbot root /var/www/grades-ui; server_name www.mygrades.co.uk mygrades.co.uk; ssl_certificate /etc/letsencrypt/live/mygrades.co.uk/fullchain.pem; # managed by Certbot ssl_certificate_key /etc/letsencrypt/live/mygrades.co.uk/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 location / { index index.html; add_header X-debug-message "Location / reached" always; } location ^~ /reset_password/(.*)[a-z]/ { alias /var/www/grades-ui; add_header X-debug-message "Location /reset_password/$1 reached" always; } location ^~ /reset_password { alias /var/www/grades-ui; add_header X-debug-message "Location /reset_password reached" always; } # location /reset_password/([0-9a-z]*) { # root /var/www/grades-ui; # try_files /index.html =404; # } } server { listen 80; if ($host = mygrades.co.uk) { return 301 $host$request_uri; } # managed by Certbot server_name www.mygrades.co.uk mygrades.co.uk; # return 301 https://$host$request_uri; # managed by Certbot proxy_set_header Host $host; proxy_set_header X-Real-IP $remote_addr; proxy_set_header REMOTE-HOST $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; }

Try: location /reset_password { root /var/www/grades-ui; try_files /index.html =404; } — Richard Smith
– Richard Smith, Commented Oct 6, 2019 at 19:12
It actually worked because it loads correctly, but the url becomes e.g. https://mygrades.co.uk/reset_password/test/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/index.html/ — curiousdev
– curiousdev, Commented Oct 6, 2019 at 19:55
To add to this: it turns out my browser was simply caching the previous results and it actually works just fine! @RichardSmith if you were to post this as the answer I will accept it. — curiousdev
– curiousdev, Commented Oct 7, 2019 at 16:18

Richard Smith · Accepted Answer · 2019-10-07 16:58:30Z

You are confusing your location syntax. The regular expression location is introduced with the ~ or ~* modifiers. The ^~ modifier means something else entirely. See this document for details.

The location ^~ /reset_password { ... } block already matches any URI that begins with /reset_password, including /reset_password/foo. You do not need a regular expression to match the trailing part of the URI.

You could use an alias directive with a regular expression location, but in that case, alias requires the entire pathname to be provided. See this document for details. For example, this should work:

location ~* ^/reset_password { alias /var/www/grades-ui/index.html; }

But the solution I would suggest is a prefix location with a root and try_files directive. See this document for details. For example:

location ^~ /reset_password { root /var/www/grades-ui; try_files /index.html =404; }

Cole Tierney · Accepted Answer · 2019-10-06 21:22:24Z

The ^~ string is a prefix location modifier which tells nginx to skip processing of regular expression locations if the prefix location is matched. A regular expression location uses just ~ or ~* to ignore case. Check the documentation for more details.

That said, if you don't need to capture any regex pattern matches, just go with a simple prefix location. Regex locations are expensive and they interfere with the natural flow of nginx. The author of nginx, Igor Sysoev, explains in this video.

A good way to use a regex location is to wrap it in a prefix location. This isolates it from the rest of the configuration and makes it easier to debug. Your config will also scale more smoothly. Here's an example:

location /reset_password/ { location ~ /reset_password/([0-9a-z]*) { # ... } }

Thank you for the detailed explanation. I have actually changed my code to use this method instead!

Collectives™ on Stack Overflow

NGINX regexp at the end of location path

2 Answers 2

Comments

2 Comments

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Related