Hi,
Its a fail over for the vpn connection. In the client configuration it has remote options for each of the ip addresses, along with the float option. The vpn attempts to connect to the first listed ip, and then the second ip if it cannot connect to the first. If the connection drops then the client vpn will automatically connect to the other remote ip (after the timeout).
Route:
0: from all lookup local
10000: from all fwmark 0x1/0xff lookup T1
10001: from all fwmark 0x2/0xff lookup T6
20000: from x.x.x.x lookup T1
20000: from y.y.y.y lookup T6
32766: from all lookup main
32767: from all lookup default
iptables (mangle table prerouting chain):
MARK all -- eth0 * 0.0.0.0/0 0.0.0.0/0 MARK set 0x1
MARK all -- eth1 * 0.0.0.0/0 0.0.0.0/0 MARK set 0x2
Table T1 and T6 route traffic out the interface associated with each connection.
I found out that responses to the keep-alive pings were switching interfaces by using a tcpdump (filtered on port 1194) on both interfaces.
I can provide more info, but I am a little stumped were to look from here, or what else to provide. When I run openvpn with increased verbosity, I can see the ping request and responses. Then when the replies switch interfaces the server responses are refused by the client untill the connection gets reinitialized on the second remote ip.