Sorry, no idea about Oracle, but: Mon Dec 13 20:09:51 2010 : Error: There appears to be another RADIUS server running on the authentication port 1812 Seems the radius proces is still running, so does not stops properly. Maybe the former rasiusd process was hanged. Do an <xx>radius stop & look via "ps -eaf |grep -ie radius" if still appears, kill it by hand. Restart radius. Hope it helps a bit. -----Original Message----- From: freeradius-users-bounces+escriba=cells.es@lists.freeradius.org [mailto:freeradius-users-bounces+escriba=cells.es@lists.freeradius.org] On Behalf Of ????????? ?????? Sent: miércoles, 15 de diciembre de 2010 9:06 To: freeradius-users@lists.freeradius.org Subject: FreeRADIUS - no service! Hi to all! We have an issue with a FreeRADIUS server (FreeRADIUS server and Oracle Database). Sometimes it fails and stops to process the AAA requests (only complete server reboot helps)! Here is the log output during the time of a failed situation: [root@aaa0 ~]# tailf /usr/local/var/log/radius/radius.log Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: Oracle logon failed: 'rlm_sql_oracle: no connection to db' Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct) in sql_accounting: stop packet with zero session length. [user 'mobile', nas '10.115.66.5'] Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct) in sql_accounting: stop packet with zero session length. [user 'mobile', nas '10.115.66.5'] Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: Oracle logon failed: '(null)' Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct): Failed to connect DB handle #7 Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct): reconnect failed, database down? Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct): Couldn't insert SQL accounting START record - rlm_sql_oracle: no connection to db Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct) in sql_accounting: stop packet with zero session length. [user 'mobile', nas '10.115.66.5'] Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: execute query failed in sql_query: ORA-00001: unique constraint (RADIUS.ACCT_IDX1) violated Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: Oracle logon failed: 'rlm_sql_oracle: no connection to db' [root@aaa0 ~]# [root@aaa0 ~]# /etc/init.d/rc.radiusd restart Starting FreeRADIUS:Mon Dec 13 20:09:51 2010 : Info: Starting - reading configuration files ... radiusd [root@aaa0 ~]# tailf /usr/local/var/log/radius/radius.log Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct) in sql_accounting: stop packet with zero session length. [user 'mobile', nas '10.115.66.5'] Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: Oracle logon failed: '(null)' Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct): Failed to connect DB handle #7 Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct): reconnect failed, database down? Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct): Couldn't insert SQL accounting START record - rlm_sql_oracle: no connection to db Mon Dec 13 19:25:12 2010 : Error: rlm_sql (sqlacct) in sql_accounting: stop packet with zero session length. [user 'mobile', nas '10.115.66.5'] Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: execute query failed in sql_query: ORA-00001: unique constraint (RADIUS.ACCT_IDX1) violated Mon Dec 13 19:25:12 2010 : Error: rlm_sql_oracle: Oracle logon failed: 'rlm_sql_oracle: no connection to db' Mon Dec 13 20:09:51 2010 : Info: Using deprecated naslist file. Support for this will go away soon. Mon Dec 13 20:09:51 2010 : Error: There appears to be another RADIUS server running on the authentication port 1812 [root@aaa0 ~]# reboot Please, advise what can be the reason of such negative behaviour of the RADIUS server? Why does it lose connection to a db? How can we troubleshoot the issue further? Alexander - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Hi,

We have an issue with a FreeRADIUS server (FreeRADIUS server and Oracle Database). Sometimes it fails and stops to process the AAA requests (only complete server reboot helps)! Here is the log output during the time of a failed situation:

from the log 1) no connection to the DB. thats very concise 2) you are trying to send a dodgy accounting packet to the DB. dont. 3) you are sending a packet with same details - constraint violation. 4) finally, you are trying to run radius when its already running still. kill -9 radiusd then run the server. why is it dying? proably numerous reasons. what version are you running? I'd advise you do 2 things, firstly, have some extra terminal windows open and in one, run a tcmdpump of your DB connection, and in the other, run radiusd -X (ie run in full debug mode as is written in all docs when you have a problem) you might also want to put a wrapper around the accounting part so that packets with zero session length etc (do you use Cisco gear by any chance? ;-) ) dont find their way into STOP packets...which is why I asked what version you are running.. in 2.1.x, something like this will do it in the accounting {} sectin of your server, instead of just calling SQL, wrapper it: if (Acct-Session-Time != 0) { sql } else { ok } alan