"Luiz Lima" <[EMAIL PROTECTED]> wrote: > I'm 95% sure about the cause and it's that the telco screws up, sending > packets from the same port-id for different customers and with different > session-ids for Start and Stop packets for the same connection (among other > stuff). > > Problem is, as in most of the places in the world, talking to the telco is > even easy sometimes, but it's nearly impossible to convince them that > they're wrong.
That does happen... > FreeRADIUS "recovers" from these glitches but by other server doesn't. > That's why my first idea was that FreeRADIUS could keep its target RADIUS > server informed of online users through Alive packets. Since it already may > be configured to have its own internal schedule for retries, I figured "why > not sync remote servers with Alive packets". The retries sent by the server are ONLY for about 30 seconds. What you're talking about is having the server send Alive packets for as long as the user is logged in, which may be hours. That means the server has probably 100 to 1000 times as much information to keep track of, which isn't a good thing. Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
