>I think you should start by determining why you are getting duplicate accounting
>records. It seems to me you should only get one start and one stop for each
>event, and that would improve the situation greatly. Check your NAS radius
>timeouts and verify in a trace 4 that Radiator is indeed sending the correct
>responses within the NAS timeout period. Then, consulting the same trace 4
>debug from Radiator, check the timestamps on the accounting events (starts and
>stops) and verify the length of time the SQL server is taking on each query -
>this will show you immediately where the time is being spent.
This is what we done. The problem is about a bug in Cisco AS5300. ;-))
>From the error message above, it would appear that you are trying to alter the
>socket read queue length in Radiator with the "SocketQueueLength" parameter in
>the configuration file? Note that this is the queue length for the socket(s)
>upon which Radiator receives radius requests from your NAS(s), it has no effect
>on your SQL database connection. Also, the NAS(s) do not open the auth and acct
>sockets, they just send to the designated UDP port numbers on the radius host
>that is specified in the NAS internal configuration.
It's true, that we put it on config file. Should we remove it?
>If you do discover from the investigations above that your SQL server is
>causing the problems because it is slow, I think you will need to address that
>problem as a seperate exercise.
Thanks
Lutfi
>hth
>Hugh
--
Radiator: the most portable, flexible and configurable RADIUS server
anywhere. SQL, proxy, DBM, files, LDAP, NIS+, password, NT, Emerald,
Platypus, Freeside, Interbiller, TACACS+, PAM, external, etc, etc.
Available on Unix, Linux, FreeBSD, Windows 95/98/2000, NT, MacOS X.
