1.Troubleshooting MySQL Connections , 2. troubleshooting possible memory leak

[Norbert Wegener]

1 .Maybe max_connections in the MySQL config file should also be 
increased and Mysql be restarted.
2. No idea except upgrading.

Norbert Wegener

Stefan A. schrieb:
> Gurus,
>
> in my radiusd.log, I can see lots of these errors:
> Wed Sep 24 09:40:54 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:40:55 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:40:56 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:40:57 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:40:58 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:41:01 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:48:00 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:48:01 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:48:02 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:48:03 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:48:04 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:49:37 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:49:38 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
> Wed Sep 24 09:51:39 2008 : Info: rlm_sql (sql_accounting): There are no DB
> handles to use! skipped 0, tried to connect 0
>
> I thougt, giving the system more DB connections and lowering the retry delay
> should help.
> It did, and I now get only about 50% of the errors. But what would be a
> reasonable value for this?
>
> 'sql_accounting' is configured:
> sql sql_accounting{
>   database = "mysql"
>   driver = "rlm_sql_${database}"
>   server = "db-master"
>   port="3306"
>   login = "yyyyy"
>   password = "xxxxx"
>   radius_db = "radius_te"
>   acct_table1 = "radacct"
>   acct_table2 = "radacct"
>   postauth_table = "radpostauth"
>   authcheck_table = "radcheck"
>   authreply_table = "radreply"
>   groupcheck_table = "radgroupcheck"
>   groupreply_table = "radgroupreply"
>   usergroup_table = "radusergroup"
>   deletestalesessions = yes
>   sqltrace = yes
>   sqltracefile = ${logdir}/sqltrace.sql
>   num_sql_socks = 8
>   connect_failure_retry_delay = 30
>   #readclients = yes
>   nas_table = "nas"
>   $INCLUDE sql/${database}/dialup.conf
> }
>
>
>
> An other issue:
> FR is eating Memory and I do not know how to troubleshoot this.
> It takes about 1 MB/ Minute which was about 3.5 GB over some Days, before we
> rcognized this.
> I configured the option to die a server after 500 packets, but this does not
> help for me. Do I have to restart the server from time to time?
> How may I troubleshoot this?
>
> May I unconfigure radutmp? I think I do not need this.
> Do I really need the sqltrace-file? How my I unconfigure it? Just deleting
> the config line
>
>
>
> My system runs FR 2.0.5 on Solaris 10 using MySQL 5.0.51a
>
> Any Ideas?
>
> Thank You.
> Regards Stefan
>
>
> -
> List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
>