So I have a funny problem. I currently have a large Drupal site that I am migrating to AWS. All functions work except one, and I can no for the life of me figure out whats broken. The current config looks like this:

                             memcached
                                   |
ELB ---> Varnish ---> Drupal (with php-fpm) ---> RDS (mysql)
                                   |
                                 solr

Client connections via HTTP and HTTPS get terminated on the ELB. This is working okay as far as I can tell.
The ELB then passes the connection to the Varnish cache server as HTTP on port 6081. Again no problems.
Varnish then connects to either Drupal on port 81 OR cacherouter.php on port 82. I am not sure if cacherouter is working properly, but I do see connections going to it. The Connections to Drupal on port 81 also work fine.
Memcache and Solr are also fine.

My problem is that there is a login page that must use SSL. Currently I think the nginx configs that handle that service catch connections going to /login and force a 302 to HTTPS. This should be okay as the ELB listens to that for that port. However I get a timeout error waiting for the page to render.

Any ideas? I can post configs if needed.

R