On Mon, 2005-06-20 at 18:54 +0800, ??? wrote:

Hello everybody,

I am using freeradius-1.0.4.

In my situation, user "testuser" from domain "domain" logged from NAS 192.168.1.68,

After a while,user "testuser" logged out,but the NAS did not send Accout-Stop packet(for some reason) and freeradius still

thought that user "testuser" is alive,but Actually user "testuser" has logged off.

I have set the Simulate-Use to 1 and now "testuser" can not log in any more from anywhere.

I use radwho and it output as follows:

Rather than just telling you to read the Documentation...I had a problem with this and feel the documentation could do with some refinement. First: Look at:- /usr/src/freeradius-1.0.4/doc/Simultaneous-Use With flat files: logged-in users are in the 'radutmp' file (something like /usr/local/var/log/radius/radutmp). 'radwho' prints the file. If 'checkrad' is set up to run, it should 'snmp' (or however its been told to run) the NAS where the user was last seen - and if the NAS says the user is no longer there, clear the radutmp entry and allow login. With MySQL (which I use): logged-in users are determined by SQL asking the accounting table for the user where the 'stoptime' is Zero (ie - no stop record received). The SQL looks like: "SELECT COUNT(*) FROM ${acct_table1} WHERE UserName='%{SQL-User-Name}' AND AcctStopTime = 0" If 'checkrad' is programmed to run, It runs the SQL: "SELECT RadAcctId, AcctSessionId, UserName, NASIPAddress, NASPortId, FramedIPAddress, CallingStationId, FramedPr otocol FROM ${acct_table1} WHERE UserName='%{SQL-User-Name}' AND AcctStopTime = 0" ... which gives it the NAS to go and ask. Whether you use radutmp or SQL for checking Simultaneous-Use is also determined by the setting for 'session' in 'radiusd.conf'. Mine looks like: session { # radutmp <-- commented out. sql } -- . . ___. .__ Posix Systems - Sth Africa. e.164 VOIP ready /| /| / /__ mje@posix.co.za - Mark J Elkins, Cisco CCIE / |/ |ARK \_/ /__ LKINS Tel: +27 12 807 0590 Cell: +27 82 601 0496