|
On Sat, 28 Sep 2002, Bjoern Metzdorf wrote:
> The tunneling solution would be the most elegant one, but it requires
> additional tunneling setup on the RIPs (although, if you use TUN or DR, you
> already have dealt with a special additional setup :-)
In my situation the tunneling setup should be less of a hassle and work
better than binding to multiple IPs/ports, so I'd like to take that route
if possible! :)
I have the tunnel up and working but I'm running into trouble actually
sending traffic to the VIP through the IPIP tunnel since the VIP is also
configured on localhost:
director# ip addr
<snip>
3: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 100
link/ether 00:03:47:7c:42:ab brd ff:ff:ff:ff:ff:ff
inet 66.33.206.13/24 brd 66.33.206.255 scope global eth1 <--
RIP
inet 66.33.206.41/24 brd 66.33.206.255 scope global secondary eth1:10 <--
VIP
<snip>
16: tunl1@NONE: <POINTOPOINT,NOARP,UP> mtu 1480 qdisc noqueue
link/ipip 0.0.0.0 peer 66.33.206.14
inet 192.168.1.1/24 scope global tunl1
director# ip route ls
66.33.206.41 via 192.168.1.2 dev tunl1 <-- send thru tunnel!
192.168.1.0/24 dev tunl1 proto kernel scope link src 192.168.1.1
66.33.206.0/24 dev eth1 proto kernel scope link src 66.33.206.13
default via 66.33.206.1 dev eth1
...but if this is going to work I'm assuming I need some special flags
when I add the route in addition to
director# ip route add 66.33.206.41 via 192.168.1.2 dev tunl1
The above obviously doesn't work since the VIP 66.33.206.41 is configured
locally on eth1:10. Is it actually possible to get around this?
Thanks!
sage
|
