Is there a Nginx environment variable equivalent to Apaches' {ENV:REDIRECT_STATUS}? - .htaccess

Question

When migrating from Apache to Nginx certain rules from the .htaccess file have to be 'translated' to Nginx the configuration file.
One problem I can't seem to solve, an example is the easiest way to explain:
The request http://www.domain.com/nginx is internally rewritten by Apache to index.php?option=com_content&view=article&id=145
Now I would like to block direct requests to index.php?option=com_content so the page is only available thru http://www.domain.com/nginx in order to avoid duplicate content. In Apache this was achieved by using these .htaccess rules:
# Check if it's the first pass to prevent a loop. In case of first pass, the environment variable contains nothing
# If http://www.domain.com/nginx already internally has been rewritten to index.php?option=com_content&view=article&id=145 {ENV:REDIRECT_STATUS} contains '200' and the request is allowed to be processed
RewriteCond %{ENV:REDIRECT_STATUS} ^$
# Check if the query string contains requests for the page
RewriteCond %{QUERY_STRING} ^index.php?option=com_content&view=article&id=145 [NC]
# If conditions apply, reject request
RewriteRule .* 404 [L]
In Nginx, is there such an environment variable that I may use?
Or should I approach this in a totally different way?
EDIT 1:
In real life, it is not only that one page but a dynamic Joomla site with lots of pages. I tested the above which works, but the intention is to block ALL requests on index.php?option_content&view=article&id=*
EDIT 2:
This is the working NGINX configuration file:
server {
listen 80 default_server;
listen [::]:80 default_server ipv6only=on;
root /usr/share/nginx/html/domainname;
index index.php index.html index.htm;
server_name domainname.com;
server_name localhost;
location / {
try_files $uri $uri/ /index.php?$args;
}
# deny running scripts inside writable directories
location ~* /(images|cache|media|logs|tmp)/.*\.(php|pl|py|jsp|asp|sh|cgi)$ {
return 403;
error_page 403 /403_error.html;
}
## give 404 header & redirect to custom errorpage without changing URL ##
error_page 404 = /404_custom.php; #global error page, script handles header
error_page 500 502 503 504 /50x.html;
location =/index.php {
set $arg_set "${arg_option}___${arg_view}___${arg_id}";
if ($arg_set ~* "^(((\w|-)+?)___){2}((\w|-)+?)$") {
return 404;
}
try_files $uri =404;
fastcgi_split_path_info ^(.+\.php)(/.+)$;
fastcgi_pass unix:/var/run/php5-fpm.sock;
fastcgi_index index.php;
include fastcgi_params;
}
location ~ \.php$ {
try_files $uri =404;
fastcgi_split_path_info ^(.+\.php)(/.+)$;
fastcgi_pass unix:/var/run/php5-fpm.sock;
fastcgi_index index.php;
include fastcgi_params;
}
}

Answer 1

The Apache approach will not work here, but there are many other ways you can go about this problem, depending on how many such rules you are going to implement and some other conditions. In general case, I would use something like this:
map "${arg_option}___${arg_view}___${arg_id}" $show404 {
default 0;
# Put here the argument value sets of the pages
# you want to hide - one set per line
"com_content___article___145" 1;
}
server {
...
location /nginx {
rewrite ^.*$ /index.php?option=com_content&view=article&id=145 break;
proxy_pass ...
}
location =/index.php {
if ($show404) {
return 404;
}
proxy_pass ...;
}
...
}
EDIT:
If you want to block all requests to index.php where the arguments "option", "view" and "id" are present, no matter what their values are, you could use something like this:
location =/index.php {
set $arg_set "${arg_option}___${arg_view}___${arg_id}";
if ($arg_set ~* "^(((\w|-)+?)___){2}((\w|-)+?)$") {
return 404;
}
proxy_pass ...
}
In case some values of these arguments should be checked, just modify the regular expression to suit your purpuse:
location =/index.php {
set $arg_set "${arg_option}___${arg_view}___${arg_id}";
if ($arg_set ~* "^com_content___article___(\d+)$") {
return 404;
}
proxy_pass ...
}
Also, in your situation map can be used to simplify the configuration, so that you wouldn't have to add another location for every article, instead encapsulating all the rewrite rules in one map block like this:
map "$request_uri" $real_args {
default "";
"~*^/nginx" option=com_content&view=article&id=145;
"~*^/some_article" option=com_content&view=news&id=123;
"~*^/another_article" option=com_content&view=article&id=515;
}
server {
...
location / {
if ($real_args) {
rewrite ^.*$ /index.php?$real_args break;
}
proxy_pass ...
}
location =/index.php {
# See above
}
...
}
EDIT 2:
For one or two exceptions you could improve your regular expression, using a negative look-ahead:
if ($arg_set ~* "^(((\w|-)+?)___){2}((?!175$)(\w|-)+?)$") {
return 404;
}
But if you expect to have a lot of such URLs, you'll have to introduce map into your configuration after all. Otherwise your regular expression will get too complicated and unmanageable. The configuration in this case would look like this:
map "${arg_option}___${arg_view}___${arg_id}" $exception {
default 0;
"com_content___article___175" 1;
"com_content___news___188" 1;
"something___else___211" 1;
}
server {
...
location =/index.php {
set $arg_set "${arg_option}___${arg_view}___${arg_id}";
if ($exception) {
break;
}
if ($arg_set ~* "^(((\w|-)+?)___){2}((\w|-)+?)$") {
return 404;
}
proxy_pass ...;
}
...
}
This might seem somewhat counterintuitive but that's just how "if" works in Nginx. The second "if" will not be evaluated if Nginx meets break in the first "if" block.