A packet arriving from the upstream proxy contains this:-
HTTP/1.0 100 Continue
Date: Thu, 16 Dec 2004 18:58:28 GMT
Via: 1.1 cache0 (NetCache NetApp/5.3.1R2)
Henrik Nordstrom wrote:
And HTTP forbids this message to be sent to Squid (MUST NOT) as Squid
is a
HTTP/1.0 client (RFC 2616
the upstream proxy contains this:-
HTTP/1.0 100 Continue
Date: Thu, 16 Dec 2004 18:58:28 GMT
Via: 1.1 cache0 (NetCache NetApp/5.3.1R2)
To which the squid server adds:-
X-Cache: MISS from jfc
Proxy-Connection: close
As it passes it back to the client.
Or am I missing the point
regards
Martyn
what is not getting through? I've tried various levels of
debugging, but can make little sense of it.
regards
Martyn Bright
TRML
01455 850444
The line from the config file is:-
cache_peer proxy.easynet.net parent 3128 0
OK, but if the ISP parent is the only one
with Internet access , and your local squid hasn't; then you need
never_direct allow all
Yes, I have that.
The access.log entries from the application
Instead of permitting only the safe ports try to let every port
through
for
that particular client and see if that helps. If it works than you
know it
is some extra ports besides the safe_ports which you have to open for
that
particular client.
I added
acl Safe_ports port 1-65535
but
,
the client has thrown its toys out of the cot.
Mow this is mostly gobbledy gook to me, but I am sure the answer is in
there somewhere.
regards
Martyn Bright
Chris Perreault wrote:
There was a patch, just mentioned recently, that was posted here about a
month ago. We had it created for us to do the following:
In reverse proxy (accelerated mode):
User wants to hit internal webserver.
Browser sends the user to the proxy which uses a
Henrik Nordstrom wrote:-
Not really. If IE understands this page contains a password form it still
allows you to save the password...
I am have an example of a Web login pages that doesn't offer you the chance
to save the password... I accept I need to be a little obscure in the way I
ask
Gareth Evans wrote
I don't want to be pushy but can anyone else help with this problem?
Thanks,
Gareth.
It's a dumb question, but your /etc/passwd.squid file does contain the
passwords in the appropriately encrypted form. Copying it from /etc/shadow
doesn't work.
exhibited
this https/auth/domain name stripping problem has been Squid. Does that
mean that the others are effectively all using non persistent connections,
or is Squid missing a trick somewhere?
Martyn Bright
that it works and what steps I will require to configure it.
Thanks
Martyn Bright
-run - NONE/- text/html
-cut-
I've tried lots of different https sites and the result is always the same.
Suggestions please!
regards
Martyn Bright
/- text/html
-cut-
I've tried lots of different https sites and the result is always the same.
Suggestions please!
regards
Martyn Bright
control and exclusively https. Regrettably
it is also the only site that the approved users ever go to, so they always
have a shortcut to it on their desktop.
Christoph Haas wrote:
On Fri, Sep 17, 2004 at 07:23:30PM +0100, Martyn Bright wrote:
That's a very annoying IE bug. If the first page
Henrik Nordstrom wrote:
On Fri, 17 Sep 2004, Martyn Bright wrote:
Unfortunately, if I try to go to any https site, the challenge response
seems to work ok, but I get a 400 Invalid URL message because something
seems to have stripped the host name from the URL. If I refresh the
screen
it to work on 1000 clients
on
100 sites all over the country and I would prefer not to have to
reconfigure
every one!
Ask your Microsoft support channel.
This is controlled entirely within the browser.
Is there an authentication method that is not controlled by the browser?
Martyn Bright
-Original Message-
Martyn Bright wrote:
Your browser is broken and forgets to set up SSL after the proxy
authentication challenge.
You may be able to work around this broken browser by setting
client_persistent_connections off, but it will cost you performance
17 matches
Mail list logo
