Nginx - Redirect Domain Trailing Dot - nginx

How can I redirect "http://domain.com." to "http://domain.com" with Nginx?
What's the recommended way of doing this? Regex or is there any other options?

The following snippet does this in a general way, without having to hard code any hostnames (useful if your server config handles requests for multiple domains). Add this inside any server definition that you need to.
if ($http_host ~ "\.$" ){
rewrite ^(.*) $scheme://$host$1 permanent;
}
This takes advantage of the fact (pointed out by Igor Sysoev) that $host has the trailing dot removed, while $http_host doesn't; so we can match the dot in $http_host and automatically use $host for the redirect.

You will need to use Regex.
server {
listen 80;
server_name domain.com.WHATEVER, domain.com.WHATEVER-2, domain.com.WHATEVER-3;
rewrite ^ $scheme://domain.com$request_uri? permanent;
}
From: http://wiki.nginx.org/HttpRewriteModule
redirect - returns temporary redirect with code 302; it is used if the substituting line begins with http://
permanent - returns permanent redirect with code 301

Related

What's the rewrite syntax for an NGINX regex location?

Using NGINX as a load balancer running on 10.1.2.15:9002, I have a need to rewrite http://10.1.2.15:9002/proxy.stream?opt=1 to http://10.1.2.15:9002/app/proxy.stream?opt=1.
Following are bits from my nginx.conf file:
http {
upstream app_cluster {
server 10.1.2.23:8080;
server 10.1.2.25:8080;
}
server {
listen 9002 default_server;
location /app/ {
proxy_pass http://app_cluster/;
}
location ~ ^/proxy.stream(.*)$ {
rewrite ^(.*)$ /app/$request_uri last;
}
}
}
By the way, I can replace the rewrite line with return 401 (for example), and I can see the 401 HTTP status returned using Chrome Developer Tools, so I know the regex is matching. I just can't get the URI rewritten properly. In fact, I only see the original request with a 406 status in Developer Tools, so I suspect something is wrong with my rewrite syntax.
Does anyone see what is wrong with this configuration?
Using $request_uri in the replacement string of a rewrite statement is problematic, as it has not been normalised and also contains the query string, which by default, rewrite will append again.
Also, your replacement string contains //, as you are appending a URI which already has a leading /.
The regular expression location is not necessary, as a prefix or exact match location will suffice and is more efficient for nginx to process. See this document for more.
For example:
location /proxy.stream {
rewrite ^ /app$uri last;
}
Make use of the matching part from the regex instead of $request_uri
rewrite ^(.*)$ /app/$1 last;

Nginx: Add subpath cause infinite redirect loop

Hi I want to add a subpath to my current app url and redirect all request to it, example:
https://example.com/ -> https://example.com/main
https://example.com/faq -> https://example.com/main/faq
server {
listen 80 default;
server_name example.com;
rewrite ^/(?!main).*$ /main/$1;
}
I try to use regex to avoid the infinite loop on itself, but it still redirect, not sure what went wrong
You are missing a capture for $1 and you should make the rewrite issue a 3xx response if you want to see the result:
rewrite ^/(?!main)(.*)$ /main/$1 redirect;
See this document for more.

Chage the part of the URL using nginx

I m using nginx webserver.
I want to change the url before it hits the server from
https://www.example.com/abc/contact-us
to
https://www.example.com/#/contact-us
Thanks in advance.
For a single URI redirection, an exact match location and return statement may be most efficient:
location = /abc/contact-us {
return 301 /#/contact-us;
}
To redirect all URIs beginning with /abc use a rewrite directive:
location ^~ /abc/ {
rewrite ^/abc(.*)$ /#$1 permanent;
}
The location block is largely redundant, but means nginx only looks at the regular expression when it needs to. See this document for more.

How to use nginx if's param $1 in rewrite statement

I have this working code in nginx config:
if ($http_host ~* ^www\.(.+)$) {
set $host2 $1;
rewrite (.*) http://$host2$1;
}
I think that string set $host2 $1; may be omitted and $1 used in rewrite statement without defining some variables. But rewrite has own $1..$9 params.
How I may use $1 form if in the rewrite statement?
I think the regex dollar forms only apply to the most recent regular expression. So you cannot combine the $1 of the if with the $1 of the rewrite without using set. However, there are simpler solutions for your scenario.
Firstly, if you know the host name (for example example.com), you can do the following:
server {
server_name www.example.com;
return 301 $scheme://example.com$request_uri;
}
server { server_name example.com; ... }
On the other hand, if you don't have a specific host name in mind, you can do the following catch-all solution:
server {
server_name ~^www\.(?<domain>.+)$;
return 301 $scheme://$domain$request_uri;
}
server { server_name _; ... }
You can find out more about this second form here.
I don't recommend catch-all solutions because it is only meaningful to have at most one catch-all server block. If possible, use the named server solution.
Also, note that you can achieve the above redirection using the rewrite ^ destination permanent; form. All these solutions avoid using the poorly regarded if directive.

Nginx rewrite syntax

I've seen a bunch of ngnix rewrites that have syntax like this:
server {
server_name www.example.com;
rewrite ^(.*) http://example.com$1 permanent;
}
I don't understand the ^(.*) part. Does the ^ take everything after the TLD of the uri?
The ^ does indeed match at the beginning of the string. In the case of nginx's rewrite directive this means the beginning of the path component of the actual URI. Unfortunately nginx's documentation is slightly incorrect. Quoting from http://www.nginx.org/en/docs/http/ngx_http_rewrite_module.html#rewrite :
If the specified regular expression matches a URI, the URI is changed as specified in the replacement string.
However, this is technically wrong. rewrite does not match the whole URI/URL but only its path component (which always starts with a / even if the user only enters e.g. http://www.example.com instead of http://www.example.com/). Therefore rewrite ^(.*) http://example.com$1 permanent; does not turn into http://example.comwww.example.com.
If I remember it correctly, the ^ just sets the Regex rule to match the start of the string.
The parentheses are used to extract that part with the $1-9 variables.
Another solution from the Nginx wiki. Link
server {
server_name www.example.com;
rewrite ^ http://example.com$request_uri? permanent;
}

Resources