Duplicate "object" definition

[Conrad Classen]

The radius config parser will not load a duplicate named routine.

You would need to change the name and then add a call to it either 
before or after the original routine call/s.

However, if you are looking to see if a User-Name attribute exists, and 
then reject it if it does, your routine, when added will do it.

Duplicate checking is handled in the sites-available configs, which you 
can either enable or disable.

If you are using a database, then the sql code is included and can be used.

Please read the instructions found in the config files for guidance.

Thanks


On 2024/11/03 23:05, Alex Zetaeffesse wrote:
> Hi all,
>
> I apologize if I don't use the right terminology.
> I'm playing/testing with 3.2.6 to better understand how freeradius behaves
> when finding duplicate definitions.
> I created a copy of the filter policy, calling it policy.d/filter2 and I
> defined the object/policy filter_username in this way:
>
> filter_username {
>          if (&User-Name) {
>                  reject
>          }
> }
>
> Now, when in the directory policy.d I only have the file filter2, I always
> get the expected Access-Reject, if I just have the default file filter I
> get Access-Accept (the user is correctly configured in the file users).
> If I have both files (filter and filter 2), when I launch radiusd -X I get,
> the file filter is loaded first and then the file filter2
>
> ...
>
> *including configuration file /etc/freeradius/policy.d/filter     <<<*including
> configuration file /etc/freeradius/policy.d/operator-name
> including configuration file /etc/freeradius/policy.d/eap
> including configuration file /etc/freeradius/policy.d/debug
> including configuration file /etc/freeradius/policy.d/accounting
>
> *including configuration file /etc/freeradius/policy.d/filter2    <<<*
> ...
>
> and I always get Access-Accept.
> How are duplicate objects treated? Is the first definition that's retained
> or the last? From my experiment it seems the first is retained.
> May you please confirm?
>
> Regards,
>
> Alex
> -
> List info/subscribe/unsubscribe? Seehttp://www.freeradius.org/list/users.html