I’ve discovered today a small bug with the load balancer when used with a PPPOE interface.
After a ethernet link down, or a PPPOE session down, the status of the gateway link stay at “Fault”.
I need to manually issue a PPPOE down in the interface list, and up.
Then the status is updating and the gateway take back his role.
I have to say that i’m using a supplementary interface ETh interface to be able to access the local IP of the modem for configuration purpose. (eth interface with an IP + nat so that we can access the modem web interface). Perhaps this does cause problems with the load balancer status system ?
You will say me wy am i using PPPOE, when we can do it in the router.
Well; because it is mandatory to use PPPOE from the router to get Ipv6, when your provider give you IPv6 as a second pppoe session.
As almost no modem on the market is IPv6 compliant, there is no other solution to get a native IPv6 connection from a consumer grade modem than controling PPPOE from the router.
Getting IPv6 from a PPPoe session is very simple, there is just a “,” to add in the pppoe config file.
Then we can do a bridge between the pppoe interface and the lan interfaces to distribute IPv6 in the network :
Create the bridge :
brctl addbr br0
ifconfig br0 up
Adding interfaces to the bridge :
brctl addif br0 eth0
brctl addif br0 ppp0
Then restrict the bridge to IPv6 :
ebtables -t broute -A BROUTING -p ! ipv6 -j DROP
Then add radvd deamon on the router to advertise the /64 or /48 IPv6 prefix your provider gave you to the LAN machines, so that they can autoconfigure with an IPv6 adress.
Anyone have a take on whether or not radvd is available in the current version (3.3.x)? If so can I push an IPv6 routed subnet out to clients, or do I need to install a major OS like Red Hat, or CentOS, possibly even Debian / Ubuntu for routing an IPv6 subnet. I would prefer not to have to create one hell of a lot of custom code to handle the Captive Portal for which I would prefer to use an off the shelf / already existing solution.