== Configuring FreeRADIUS as a RadSec test client

Unfortunately, the `radclient` program does not support RadSec.  We
must therefore configure an instance of FreeRADIUS as a "transport
converter" which proxies UDP-based RADIUS requests to a RadSec
destination of our choice.

The following steps should be performed on a client system, which we
will call `radseccli`.  This system should be a new system, with a
different IP address.  That is, you shoudl not edit the configuration
on the `radsecsvr` host.  Doing so will break the RadSec configuration.

Install FreeRADIUS using the NetworkRADIUS packages by following the
instructions provided here:

<https://networkradius.com/packages/>

Before making any configuration changes, you should stop the radiusd
service:

[source,shell]
----
 service radiusd stop
----

Add a new `tls` home server definition, which will point to the RadSec
server.  We do this by creating a file
`/etc/raddb/sites-enabled/radsec-homeserver` with the following
contents:

.Example homeserver, pool and realm definitions for the RadSec service
====

 home_server tls {
         ipaddr = 172.23.0.3    # IP address of our RadSec server
         port = 2083
         type = auth+acct
         proto = tcp
         tls {
             private_key_password = whatever
             private_key_file = ${certdir}/client.pem
             certificate_file = ${certdir}/client.pem
             ca_file = ${cadir}/ca.pem
          }
 }
 home_server_pool tls {
         type = fail-over
         home_server = tls
 }
 realm tls {
         auth_pool = tls
         acct_pool = tls
 }

====

[TIP]
====
Complete descriptions of each of the above configuration items can be found in the
`[raddb]/sites-available/tls` example configuration file.  For simple tests, however,
we can omit all of the comments from the file.
====

To use this `tls` home server, we change the `default` virtual server to proxy
all authentication and accounting requests to it.

Edit the `/etc/raddb/sites-enabled/default` file so that the beginning of
the `authorize` and `preacct` sections looks as follows:

.Example default virtual server modification to proxy requests to a RadSec proxy server
====

 authorize {
     update control {
         &Proxy-To-Realm := tls
     }
     handled
     ...
 }
 ...
 preacct {
     update control {
         &Proxy-To-Realm := tls
     }
     handled
     ...
 }

====

These changes make the `tls` virtual server always proxy packets.
These changes are only for testing, and should never be used in
production.

We must now copy the example CA certificate as well as the client
certificate and key files which are on the `radsecsrv` host to this
test client.

Replace the following files on `radseccli` with the equivalent files from
`radsecsrv`:

[cols="1,1,1"]
|===
|File|Corresponding configuration item|Purpose

|/etc/raddb/certs/ca.pem
|`ca_file`
|CA certificate which is used to authenticate the server certificate presented by the RadSec server to the client.

|/etc/raddb/certs/client.pem
|`certificate_file`
|Client certificate (signed by the CA certificate) that is presented by the test client to the RadSec server.

|/etc/raddb/certs/client.pem
|`private_key_file` and `private_key_password`
|Private key corresponding to the client certificate
|===

Note that the client certificate and key are typically bundled into a single file.

[CAUTION]
====
If you do not correctly replace the CA, client certificate, and key
material on the test client then the RadSec client and RadSec server
will fail to mutually authenticate each other as they do not share a
trusted CA.  If you see messages like `unknown CA`, then you know that
the certificates have not been set up correctly.
====

Start the FreeRADIUS service in debug mode:

[source,shell]
----
radiusd -X
----


=== Testing RadSec connectivity

At this stage you should be able to cause the test client to send RadSec
requests directly to the RadSec server.

Run the following to send a RADUS (UDP) Access-Request to the local FreeRADIUS
instance.  It  should then proxy the request over RadSec connection to
the remote RadSec server:

[source,shell]
----
 echo "User-Name = bob" | radclient 127.0.0.1 auth testing123
----

If the test client is able to successfully establish the RadSec
connection, and the RadSec server replies with an Access-Accept
response, then the output will be as follows:

.Expected output from radclient
===============================

 Sent Access-Request Id 252 from 0.0.0.0:50118 to 127.0.0.1:1812 length 27
 Received Access-Accept Id 252 from 127.0.0.1:1812 to 127.0.0.1:50118 length 39

===============================

Lack of response or an Access-Reject response indicates that the RadSec
connection is not being established successfully.

There may be serveral reasons for broken connectivity including:

  * The client not accepting the certificate presented by the server.
  * The server not accepting the certificate presented by the client.

Look at the debug output generated by both the test client and the RadSec
server. In many cases it will tell you exactly what the problem is.

Do not proceed with any further steps until direct connections between the
RadSec client and Radsec Server are working properly.

Once things are working we are ready to
xref:protocols/proxy/radsec_with_haproxy.adoc[configure HAproxy to proxy RadSec
connections] or to xref:protocols/proxy/radsec_with_traefik.adoc[configure
Traefik to proxy RadSec connections].