manny wrote:
I am attempting to setup a FreeRADIUS server to listen only on one IP address, but have two virtual servers.
The documentation says that packets coming from a "listen" section go to one virtual server. You will need to have multiple "listen" sections.
At the moment I am testing with simple servers, one that accepts and one that rejects. I have a freeradius server that is dependent on another server, and want to provide an alternate authentication method when the server is down.
What you can do is to put the home servers into a fail-over pool. Have the proxy fail over from the external home server, to one that you control.
home_server_pool my_auth_failover { type = fail-over home_server = first home_server = second }
That should works.
This is the result from radtest
<sigh> If you're debugging the server, why are you looking at radtest, instead of the output of "radiusd -X"? This is documented in the FAQ, web page, documentation, "man" pages, and daily on this list.
[root@localhost raddb]# radtest test test123 192.168.100.251 1812 secret Sending Access-Request of id 1 to 192.168.100.251 port 1812 User-Name = "test" User-Password = "test123" NAS-IP-Address = 127.0.0.1 NAS-Port = 1812 rad_recv: Access-Reject packet from host 192.168.100.251 port 1812, id=1, length=20 [root@localhost raddb]#
It is my understanding that once a home_server is dead, it tries the next one, but here I see that port 1812 is still used instead of 1822.
You do understand that "radtest" is not a RADIUS server, right? Radtest will send packets to where *you* tell it to send packets. Here, you've told it to send packets to port 1812. So it does that. radtest is *not* FreeRADIUS. Again, look at the debug output of the *server* to see how the *server* is behaving. Alan DeKok.