Has anyone checked for Acct-Terminate-Cause = User-Error with the packets in question? We are seeing some of the "Info: [sql] stop packet with zero session length." messages logged (but only so far during 1 specific minute of the day) and they ALL have "Acct-Terminate-Cause = User-Error". There is one other packet with Acct-Session-Time = 0 AND Acct-Terminate-Cause = User-Request that does NOT generate the warning. Our current belief is that this is NOT radius related. The troublesome packets do not appear to be causing any harm. (despite the subject, radiusd is NOT exiting and remains healthy and happy (Yeah 2.1.8!! ) I'm just thinking you guys might be chasing a non (radius) issue. Cheers, -craig ----- Original Message ----- From: "Alan DeKok" <aland@deployingradius.com> To: "FreeRadius users mailing list" <freeradius-users@lists.freeradius.org> Sent: Wednesday, January 06, 2010 5:17 AM Subject: Re: Radiusd process exited without notice
Bjørn Mork wrote:
Alan DeKok <aland@deployingradius.com> writes:
The Acct-Session-Time should be at least 1 second.
Why? I can't find any such requirements in RFC 2866.
The RFC's miss a *lot*. It is permissible for a multi-homed NAS to send an accounting "start" from one IP, "update" from another IP, and "stop" from a third IP. While this is insane, there is equipment that behaves this way.
If the session lasts less than .5 seconds then an Acct-Session-Time of 0 makes perfect sense. IMHO.
Hmm... I still don't like it.
I can agree that the near simultaneous start and stop requests indicates some error, but as suggested by Dinh Pham Cong it may be a user disconnecting immediately after the session is established. That happens.
Yes.
Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
__________ Information from ESET Smart Security, version of virus signature database 4748 (20100106) __________ The message was checked by ESET Smart Security. http://www.eset.com