I've got a WordPress blog running on lighttpd-1.4.19, listening at www00:81. On a single host, apache-2.2.11 listens on port 80, which produces a proxy connection from http://blog.mydomain.org:80 to http://blog.mydomain.org:81 (both being exactly the same physical machine).

The Apache virtualhost looks the following:

<VirtualHost *:80>
    ServerName blog.mydomain.org
    ProxyRequests Off

    <Proxy *>
    Order deny,allow
    Allow from all
    </Proxy>

    ProxyPass / http://blog.mydomain.org:81/
    ProxyPassReverse / http://blog.mydomain.org:81/
</VirtualHost>

Using debug.log-request-handling = "enable", I recieve the next log entry after i browse http://blog.mydomain.org:80 (spot the Host headers):

2010-05-10 08:47:14: (request.c.294) fd: 6 request-len: 853 
GET / HTTP/1.1
Host: blog.mydomain.org:81
[...]

2010-05-10 08:47:15: (request.c.294) fd: 8 request-len: 754 
GET /wp-content/uploads/2010/01/image.gif?w=280 HTTP/1.1
Host: www00:81

My problem: so far as I understand, the PHP atmosphere variable $_SERVER['HTTP_HOST'] is placed to that particular Host header variable. Regrettably, WordPress uses that variable within their system to produce Web addresses to pictures around the blog. These Web addresses will not be around behind a firewall obviously. How do i pressure the host header to become blog.mydomain.org rather than blog.mydomain.org:81, correspondingly www00:81? I already added set server.name = "blog.mydomain.org" to my lighttpd.conf, but this did not work.

Any suggestions are appreciated, thanks.

Perhaps you have attempted editing your [cde]? It ought to be easy to override the automated host recognition.

Namely: