new Squid3 makes many things very easy for my configs. So i tried it.
One Squid3 should be used to accel many backend domains/servers.
Some of the setups are https between browser and accel, http between accel and server.
Everything works fine if using standard portnumbers on the backend servers (80 for http, 443 for https).
My problem is, the backend server uses port 7777. It cannot be chaned to use port 80 (limitation of the backend servers software).
10.113.209.1 = squid3 accel, listening on port 7777 10.113.209.219 = backend server, listening on port 7777
Extracts from my squid.conf:
https_port 7777 cert=server.crt key=server.key defaultsite=official.site.name protocol=http
cache_peer 10.113.209.219 parent 7777 0 originserver name=tviprod
acl tviprod dstdomain official.site.name
cache_peer_access tviprod allow tviprod
cache_peer_access tviprod deny all
From /etc/hosts on accel: 10.113.209.219 tviprod
The Errormessage: 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| TCP connection to 10.113.209.219/7777 failed 2003/07/25 10:52:02| Detected DEAD Parent: tviprod
A tcpdump shows me, that squid3 actually tried to connect on port 80 instead of 7777
Is this a bug or am i doing something wrong?
Similar setups on squid2.4 work fine when using the follwoing directives, but they seem to have gone in squid3.
http_accel_with_proxy on
http_accel_uses_host_header on
Best Regards
Matthias Weigel ISOnova GmbH
