[squid-users] Strange behaviour of NTLM and "helperStatefulDefer: None available"

Dear listmember,

I somehow managed to setup squid2.5 Stable1 using ntlm/smb authentication.

The situation is like this:

I request a site.
I see the site build up in my ms-ie.
In cache.log I see the request three times. The first two times it is denied
(407), the
third time it returns 200.

This is just strange to me, but actually not an error at all.

But there are two more problems. First, if a page loads, for every
element in this page (for instance every .gif from a page), ntlm performs a
single authentication on our windows domain controller. This might be ok for

testing purposes, but Iam afraid of performance issues, if hundreds of users
accessing.

The second - and the biggest - problem is when hitting reload very often, or
accessing
some pages quite fast. Squid produces a core-dump, exiting abnormally.

cache.log says:

FATAL: Too many queued ntlmauthenticator requests

And exits. Before that, there where several:

2003/01/14 14:32:21| helperStatefulDefer: None available.

showing up in cache.log

My explanation for this so far is the following: When squid performs a login
procedure
for every object requested from web, it is by far not fast enough. It queues
up some
requests, finally dying when it has queued to much.

Is this correct? Does the login procedure have to happen for every object on
every
page? I thought ntlm does an auth once for a time-span?

Ive already searched for an option to turn on more detailed debugging on
ntlm, but
just found a debug-flag in ntlm.h.

Thanks alot in advance!!

Regards

-- 
+++ GMX - Mail, Messaging & more  http://www.gmx.net +++
NEU: Mit GMX ins Internet. Rund um die Uhr für 1 ct/ Min. surfen!