problem with bonding and IPv6
From: Jeff Haran
Date: Tue May 13 2008 - 17:34:51 EST
Hi,
When I ifenslave two ethernets under a single bonded interface, I find
that the IPv6 link local address generated for the bonded interface is
identical across multiple hosts:
On one host:
switch1:root> ip addr show dev bond0
7: bond0: <BROADCAST,MULTICAST,MASTER,UP> mtu 1500 qdisc noqueue
link/ether 00:05:1e:39:34:73 brd ff:ff:ff:ff:ff:ff
inet 10.33.72.202/20 brd 10.33.79.255 scope global bond0
inet6 2001::1235/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::200:ff:fe00:0/64 scope link
valid_lft forever preferred_lft forever
switch1:root>
On another:
switch1:root> ip addr show dev bond0
7: bond0: <BROADCAST,MULTICAST,MASTER,UP> mtu 1500 qdisc noqueue
link/ether 00:05:1e:39:ff:3a brd ff:ff:ff:ff:ff:ff
inet 10.33.72.204/20 brd 10.33.79.255 scope global bond0
inet 10.33.72.203/20 brd 10.33.79.255 scope global secondary bond0:1
inet6 2001::1236/64 scope global
valid_lft forever preferred_lft forever
inet6 2001::1234/64 scope global
valid_lft forever preferred_lft forever
inet6 fe80::200:ff:fe00:0/64 scope link
valid_lft forever preferred_lft forever
switch1:root>
This seems to negatively effect neighbor discovery on at least one of
them:
switch1:root> ip neigh show
fe80::200:ff:fe00:0 dev bond0 nud failed
fe80:60:69bc:70::2 dev bond0 lladdr 00:19:aa:2d:c4:60 router nud stale
fe80:60:69bc:70::1 dev bond0 lladdr 00:05:73:a0:00:46 router nud stale
fe80:60:69bc:70::3 dev bond0 lladdr 00:19:56:32:86:e0 router nud stale
127.1.1.8 dev eth1 lladdr 00:05:1e:39:ff:3b nud reachable
switch1:root>
These link local addresses are supposed to be unique, derived from the
MAC address of the interface, if I am not mistaken.
Is this a known problem? Are others seeing it? Is there a fix?
Kernel version 2.6.14.2
Thanks,
Jeff Haran
Brocade
--
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html