|
Hi!
I'm building an HA cluster based on RedHat/VaLinux 6.2 + piranha/ipvsadm/...
I've got two ethernet cards in each cluster's node, each of them is
connected to a different switch. For example, on a real web server, they're
configured as follows on an 192.168.1.0/24 subnet:
192.168.1.1 eth0 Link encap:Ethernet HWaddr 00:C0:DF:A0:41:D9
192.168.1.11 eth1 Link encap:Ethernet HWaddr 08:00:09:C0:02:84
I've got two real web servers and I would like to use them as they were
four, letting pulse monitor their 4 IP adresses.
I've got a problem: I've tried to ping, from another host of my subnet,
192.168.1.1 and 192.168.1.11, but everytime I ping both of them, the SAME
hardware interface replies to me. Infact, launching arp -an:
? (192.168.1.1) at 08:00:09:C0:02:84 [ether] on eth1
? (192.168.1.11) at 08:00:09:C0:02:84 [ether] on eth0
The problem is that if I disconnect the cable of eth1 interface, the server
doesn't reply anymore, even if it should have the other interface active!
The other interface starts working only if i "kill" eth1 with "ifconfig
eth1 down"
Is this normal? Is there a way to "statically bind" each mac address to an
IP address and to have some sort of high availability?
Thanks,
Pietro Ravasio
|
