Ok, here is the info :

Ascend 4060's.   Radius TimeOut in the NASs is set to 20 seconds.  (ie, if
the NAS doesn't get a reply from the radius server (Radiator of course), it
attempts it again in 20 seconds.)   (Cool...nested brackets)

However, if a customer logs off prior to 20 seconds online (ie, 15 second
connection, either failure of some sort, or quick Email check) the Stop
record is recieved BEFORE the start request.  No sessions are lost in this
way, but because the START request comes last, the session is added to the
online.db, but never removed, since the STOP has already been processed.

I was wondering whether  this is a problem with NT that would not exist
under Linux, or is this a universal problem.

I find it very hard to believe that Radiator is overloaded so much it can't
process the Requests coming into it.  Does Radiator have a buffer where it
stores incomming requests and handles them as soon as it is free too?  Or
does it just drop the information if it's not ready to process it?


Anton Sparrius
Smarter Way    http://www.smart.net.au
Email                [EMAIL PROTECTED]
Phone              (03) 9846 1711
         Melb       1800-240-829
         Sydn       1800-888-761

To unsubscribe, email '[EMAIL PROTECTED]' with
'unsubscribe radiator' in the body of the message.

Reply via email to