On 10/06/11 15:32, joanroldan wrote:
Hi everybody,
I have take a look to this post:
http://freeradius.1045715.n5.nabble.com/MSCHAP-Authentication-Issue-td2785146.html
The issue mentioned in that post was fixed in 2.1.10.
Are you running 2.1.10?
And I totally agree with the behaviours described in the table.
Do you? That table does NOT contain:
user@domain
...format credentials.
I have been experiencing the same. If I type the credentials on prompt of
Windows supplicant like
MYDOMAIN.COM\user
password
The request is accepted:
Fri Jun 10 15:58:51 2011 : Info: ++[eap] returns ok
Fri Jun 10 15:58:51 2011 : Auth: Login OK: [IRTA_NT\\jroldan/<via Auth-Type
= EAP>] (from client WLC_SSCC port 1 cli 00-26-B6-59-F1-EA)
But if I type
u...@mydomain.com
password
I get the next info:
This is not useful. Please show a full debug, from running "radiusd -X |
tee log".
As the configuration of freeradius is for eduroam purpose it would be nice
that user enter the credentials like u...@domain.com, and in function of the
domain the request would be proxied or not.
Is is possible to use this format in spite of NT-domain?
Yes, you just need to configure it appropriately. I suspect your
problems is that your "ntlm_auth" line is expanding to:
ntlm_auth --username=user@domain ...
...which won't work. %{mschap:User-Name} only handles
DOMAIN\user
host/name.domain.com
plainuser
...it does NOT handle:
user@domain
...but it maybe (probably?) should.
Since a virtual server which is handling eduroam should *ALWAYS* see
user@domain, you can write something like this in "inner-tunnel":
authorize {
if (User-Name =~ /^([^@]*)@(.+)$/) {
update request {
Stripped-User-Name := "%{1}"
Realm := "%{toupper:%{2}}"
}
}
else {
reject
}
# FIXME: you should check the Realm here is one of yours
# ...other modules
}
...and then in modules/mschap, the "ntlm_auth" config:
--username=%{Stripped-User-Name:-%{mschap:User-Name}}
...which will expand as follows:
user@domain -> --username=user
DOMAIN\user -> --username=user
user -> --username=user
-
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html