This problem is not limited to just phpMyAdmin. I am trying to do the same thing using nginx 0.7.65 (ubuntu 10.04 stable nginx package) in a vhost that uses / for a RoR application via mongrel_cluster and then I want multiple php aliases to reside under / like /opensis and /phpmyadmin, and both programs experience the same login endless loop issue...

In this case I cannot simply use a symlink because I want to run a php application in an alias folder inside the RoR application.

Here are my current location blocks including the mongrel_cluster location which is the first one:

location / {
root /var/www/sis/public;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header Host $http_host;
proxy_redirect off;
proxy_read_timeout 300;
proxy_pass http://mongrel/;
}

location ~ ^/phpmyadmin/(.*\.php)$ {
alias /var/www/phpmyadmin/$1;
fastcgi_pass 127.0.0.1:9000;
fastcgi_param HTTPS on;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
include fastcgi_params;
}

location = /phpmyadmin {
rewrite ^ $scheme://$host$uri/ permanent;
}

location /phpmyadmin/ {
index index.php;
alias /var/www/phpmyadmin/;
}

The above configuration gets me closer than anything else I have found so far. PHP is served and the default document defaults to index.php for the alias but for some reason it is still not the right configuration because of the login loop issue.

Hopefully one of the devs can respond to this with a working configuration. I think part of the problem is that everyone is using different versions of nginx and bug fixes and code changes have modified how aliases function in nginx.

Who knows, maybe my whole problem is an outdated version of nginx...