[squid-users] Problem: Squid3 accelerator does not cache

I would like to know why my SQUID3 installation does not cache anything
while configured as a http accelerator.

I am trying to use squid-3.0-PRE3 as a http-accelerator, using http_port
configuration:
   http_port 10.10.1.39:80 vhost defaultsite=10.10.3.46

It seems that all accesses are denied, unless I use also the
   always_direct allow ALL

configuration option or configure origin server as cache peer:
   cache_peer 10.10.3.46 parent 80 0 originserver

Now the squid passes traffic through, but _it_does_not_cache_anything!??

I started lynx and fetched a html page. Then I stopped the lynx and
started it again and requested the same page. In this case the squid2.5
said TCP_HIT whereas the squid3 said TCP_CLIENT_REFRESH_MISS. I really
didn't request the page reload (Shift-R). Below is the log file entry,
with mime_headers logging enabled.

1091015921.988 2 10.100.4.2 TCP_CLIENT_REFRESH_MISS/200 3268 GET
http://laulurastas2.intra.wasalab.com/ - FIRST_UP_PARENT/10.10.3.46
text/html [Host: laulurastas2.intra.wasalab.com\r\nAccept: text/html,
text/plain, text/sgml, */*;q=0.01\r\nAccept-Encoding: gzip, compress,
bzip2\r\nAccept-Language: en\r\nUser-Agent: Lynx/2.8.6dev.4 libwww-FM/2.14
SSL-MM/1.4.1 GNUTLS/0.8.12\r\n] [HTTP/1.1 200 OK\r\nDate: Wed, 28 Jul 2004
11:58:41 GMT\r\nServer: Apache/1.3.23 (Unix) (Red-Hat/Linux)
PHP/4.1.2\r\nLast-Modified: Tue, 09 Apr 2002 18:56:58 GMT\r\nETag:
"16ddb-b4a-3cb3397a"\r\nAccept-Ranges: bytes\r\nContent-Length:
2890\r\nConnection: close\r\nContent-Type: text/html\r\n\r]

Could somebody please explain me what could be matter or suggest further
tests and configuration options. Or do I really need to go one step back
and use squid2.5?

Regards

   -Petri Välisuo

-- 
Petri Välisuo / WasaLab Oy
PGP: http://www.wasalab.com/~petri/gpg_public_key.txt
ICQ: 176287479