fre 2006-05-05 klockan 13:45 +0200 skrev Tomasz Marciniak:
> First: reality check - is it possible to use single (for start) Squid as 
> acceleration cache for multiple webservers, each one having multiple 
> virtual hosts in an architecture like that:
Yes, as long as each site can be identified by Squid..
> Each of webservers has some number of apaches with virtual hosts or 
> java-based web applications operating on HTTP.
As long as they are speaking HTTP there is no problem.
> http_port               80        vport
Hmm.. not entirely OK. See the http_port description again. From your
description it sounds like you should be using vhost, and perhaps a
defaultsite as well.
vport -> IP based virtual servers
vhost -> domain based virtual servers
> redirect_rewrites_host_header   off
Don't use redirectors if you don't need to.
> http_access         allow all
Normally an accelerator/reverse proxy should only allow access to the
sites it acts as an accelerator for..
> cache_peer          webserver1.internal      parent      80          0           
> no-query originserver name=webserver1
Ok.
> cache_peer_access  webserver1              allow all
> cache_peer_domain  webserver1              webserver1.external.com
Either use cache_peer_access or cache_peer_domain. Mixing the two can be
a bit confusing..  (but works..)
> While trying to retrieve the URL: http://<private_ip_of_squid>/
> 
> The following error was encountered:
> 
> * Unable to forward this request at this time. 
Makes sense. See above.
Regards
Henrik
This archive was generated by hypermail pre-2.1.9 : Thu Jun 01 2006 - 12:00:02 MDT