Re: [Help-gnu-radius] 2 Instances Radiusd 1.1 and CLID question.

[Sergey Poznyakoff]

Hi Lewis,

> Hi, I am using 1.1 and it seems that when I first start Radiusd there is
> one instance of radiusd. After the first user authenticates I see where
> another instance of radiusd is running.

It's OK. Radiusd doesn't handle the requests itself, its subprocesses
do that. In your case:

> root    174  0.0  1.2  4588 3172  ??  I     1:06AM   0:00.62
> /usr/local/sbin/radiusd -y
> root    132  0.0  1.2  4496 3004  ??  I     1:05AM   0:00.10
> /usr/local/sbin/radiusd -y

The first radiusd (pid 132) is the dispatcher process, the second one
(pid 174) is the request hanlding subprocess. If the request handler
is busy at the time a request arrives, the dispatcher will spawn
another one and so on. The maximum number of the subprocesses
is controlled by `max-processes' statement in the `option' block
of raddb/config file. By default it is 16, so you may see
up to 17 radiusd processes in your ps output. A subprocess terminates
if it did not receive any requests within an idle timeout interval.
By default this interval is 3600 seconds. It can be set using
`process-idle-timeout' keyword in your raddb/config.

> Also; using an
> AS5350 radius.log shows funny for CLID...
> Jul 23 10:05:38 Auth.notice: (AUTHREQ AS5350 182 alyson): Login OK
> [alyson], CLID async

The word after `CLID' comes from the value of Calling-Station-Id
attribute, as sent by NAS.

Regards,
Sergey