[squid-users] Setting up reverse proxy. Incomplete docs

I have just finished installing and setting squid in reverse-proxy/web-acceleration where the web-server is on the same machine to the best of my current abilities.
I managed to get it working after a while. This was done ‘in spite’ of the documentation ☺
Some things missing from it, unless I got it all wrong and I would be grateful for a correction:
1. When httpd is listening on port 81, the port has to be added to the safe ports acl (acl Safe_ports port 81)
2. To allow traffic to squid, "http_access deny all" must be changed to "http_access allow all". (remember: we are talking of a web-acceleration mode so it’s a basic need to allow the web access, not only an option)
3. visible_hostname must be set. Squid didn’t want to start for me otherwise (is it because I forgot to do something else?)

In addition, I have encountered some problems:
I. httpd_accel_host virtual doesn't work for me. It shows me a DNS error trying to locate the IP or the requested page server. Why is that? I will definitely have virtual hosts and multiple accelerated server, so I must solve this issue right? - How?
II. I don’t want to cache anything for too long (especially not static htmls that seems to be cached with no expiration date at my browser - which is bad). I took the docs suggestion and used "httpd_accel_with_proxy off" (plus the fact I set the httpd_accel_host) BUT it doesn’t work. As I said - HTMLs are cached in my browser with no expiration date. Why? How can I make it work? Instead of it I added "acl QUERY urlpath_regex http" and that seems to work only sometimes (on .txt files etc) - I don’t like it though. What do I do?
III. Why really do I have to add 81 to the safe ports? Since squid uses it to fetch from the accelerated server, and its not really a port to which clients direct requests.

Ron Kass
Received on Tue Oct 15 2002 - 17:20:16 MDT