Unresponsive child and accounting

[Christophe Saillard]

Hi,

I use freeradius (1.1.7) to authenticate wireless users (EAP-TTLS/PAP) 
with an OpenLDAP backend.

Our first experience with Freeradius on a FreeBSD server was a nightmare 
(it seemed to be a thread related problem, the server stopped working 
with a lot of "unresponsive child" error logs).

So, we tried on a Linux server (kernel 2.6.22-14-server ubuntu feisty 
fawn)  and it worked fine since last week :

Wed Nov 21 15:33:21 2007 : Auth: Login OK: [xxxx] (from client localhost 
port 576353 cli 001c.bf09.480c)
Wed Nov 21 15:33:21 2007 : Auth: Login OK: [xxxx at u-strasbg.fr] (from 
client wds3 port 576353 cli 001c.bf09.480c)
Wed Nov 21 15:33:22 2007 : Error: WARNING: Unresponsive child (id 
3046112160) for request 2419782 (in component accounting module rlm_radutmp)
Wed Nov 21 15:33:22 2007 : Error: WARNING: Unresponsive child (id 
2841623456) for request 2419798 (in component accounting module rlm_radutmp)

The CPU went up to 100%.

There was about 300 802.1X clients connected (with a 2 minutes reauth 
period).

At this time we had no other choice than upgrading the hardware, it runs 
now on a 8 processor server but even with more CPU power we noticed a 
20% system load.

Here's the threading part of the radiusd.conf :

max_request_time = 30

delete_blocked_requests = no

cleanup_delay = 5

max_requests = 1000

thread pool {
         start_servers = 10
         max_servers = 1000
         min_spare_servers = 15
         max_spare_servers = 30
         max_requests_per_server = 300
}


I don't know if it's relevant but there were about 80 Eduroam users 
connected when the problem happens.

Thanks.

-- 
---------------------------
Christophe Saillard
Centre Réseau Communication
Université Louis Pasteur
---------------------------
Tél : 03 90 24 03 17
Fax : 03 90 24 03 12
---------------------------