Thus spake Mark Thornton
There has been a fair amount of discussion about missing radius records, specifically stop records as regards to concurrency control. I posted a question about this a week or so back and didn't get a response. In my case I have a single chassis that is prone to this problem, but the others are not. Any ideas why?
Also, if getting the stop record is so unreliable how do those of you who do hourly billing keep the system working? In our case the session is lost and not billed for. I can live with that but it bothers me that it isn't reliable.
Its not so much that the stop records are unreliable...the RADIUS standard indicates that RADIUS accounting is resent, to ensure that it gets through (unlike syslog which sends it once...if it doesn't get through...so sorry). The problem is really that if you have multiple RADIUS servers for redundancy, the start may get sent to one and the stop to the other, and its difficult to coordinate this between your RADIUS servers. Also its just the whole idea of forcing state on what's designed to be a stateless protocol (RADIUS). -- Jeff McAdams Email: jeffm@iglou.com Head Network Administrator Voice: (502) 966-3848 IgLou Internet Services (800) 436-4456 - To unsubscribe to usr-tc, send an email to "majordomo@xmission.com" with "unsubscribe usr-tc" in the body of the message. For information on digests or retrieving files and old messages send "help" to the same address. Do not use quotes in your message.