Radius server Crashing

[Peter Nixon]

On Mon 23 Jul 2007, Chris Bell wrote:
> Hello All -
>
> A newbie I am so please be patient with me.  I am a Junior Network Admin
> who has unexpectedly inherited a Senior Admin position. 

Congrats.

> I also inherited a problem with our Radius servers.

Welcome to the list :-)

> We are running freeradius-1.1.3-1.2.el5.i386.  Our Network backup daemon
> (Rancid) logs in to each network device on the hour to check for changes,
> backup the configs, etc...
>
> The radius server crashes after a period of time.  I think the Rancid
> service is crashing it.  When I log into the box running the radius
> server, and check the running processes associated with radius, I get:
>
> radiusd  21873     1  0 05:01 ?        00:00:05 /usr/sbin/radiusd -y
> radiusd  22006 21873  0 05:41 ?        00:00:00 [netstat] <defunct>
> radiusd  22115 21873  0 06:40 ?        00:00:00 [netstat] <defunct>
> radiusd  22417 21873  0 07:40 ?        00:00:00 [netstat] <defunct>
> radiusd  22612 21873  0 08:41 ?        00:00:00 [netstat] <defunct>
> radiusd  22937 21873  0 10:41 ?        00:00:00 [netstat] <defunct>
> root     23300 23197  0 12:14 pts/0    00:00:00 grep rad
>
> The radius server(s) are running on Centos 5 if that makes a difference.
> After these "defunct" processes build up to about 10 or 12... Radius no
> work. Boss get mad.

heh. They tend to do that on occasion...

> I've searched the net and found no answers... If the answer to my problem
> is painfully obvious, please remember that I started this email with a
> declaration of newbness.

It looks like your copy of radius is running netstat which is not 
exiting/being cleaned up properly. FreeRADIUS does NOT run netstat by 
default, so you probably have a custom module configured and it looks like 
it has issues. Maybe if you share your radiusd.conf we will be able to help 
you further...

Cheers
-- 

Peter Nixon
http://peternixon.net/