On Sun, Feb 24, 2008 at 2:59 PM, Ivan Kalik <
tnt@kalik.net> wrote:
>'Idle-Timeout' timeout is an argument in Radius. Chillispot use this
>radius argument.
Yes.
>I think freeradius configuration can update automatically the
>Acct-Session-Time argument if it see a user is IDLE.
>What do you think about this?
No. Chillispot does the accounting and sends data to freeradius. Radius
server has no idea about whether the user is connected or not. It is up
to the NAS to send that information to it. If Chillispot disconnects the
user and doesn't send accounting Stop packet to freeradius the session
in radius database will remain open. As far as I can see that's how you
set up the Chillispot - disconnect the session only on clicking the
logout link. Set up Chillispot properly.
Ivan Kalik
Kalik Informatika ISP
Yes, and a correct functioning chillispot definitely sends an accounting stop packet when idle-timeout is reached and user gets logged out.
I got idle-timeout in radgroupreply too, and as soon as chillispot reaches the idle-timeout for a user the user gets disconnected (by chillispot) and the accounting session is closed in radacct. In acctTerminateCause you should find idle-timeout as termination cause in this scenario. This is standard behaviour for chillispot in a correct setup. So I don't understand why this isn't working for the OP.
Sniff your network and see if an acctstop packet is sent by your chillispot when idle timeout is reached.