On Wed, 2005-06-22 at 13:56, Roland Dreier wrote: > Hal> The comment is not accurate. The description would be more > Hal> accurate to say that it "forces" the use of one of the IP > Hal> interfaces which share this source IP address (if a route > Hal> exists to the destination). In the case of IPoIB, there is > Hal> only one interface per IP address. > > Why do you think there's only one interface per IP address? For > example, you can easily try the following: > > # modprobe ib_ipoib > # ip addr add dev ib0 192.168.0.2/24 > # ip addr add dev ib1 192.168.0.2/24 > # ip addr show > [...] > 7: ib0: <BROADCAST,MULTICAST> mtu 2044 qdisc noop qlen 128 > link/[32] 00:02:04:04:fe:80:00:00:00:00:00:00:00:02:c9:01:07:fc:c7:11 > brd 00:ff:ff:ff:ff:12:40:1b:ff:ff:00:00:00:00:00:00:ff:ff:ff:ff > inet 192.168.0.2/24 scope global ib0 > 8: ib1: <BROADCAST,MULTICAST> mtu 2044 qdisc noop qlen 128 > link/[32] 00:02:04:05:fe:80:00:00:00:00:00:00:00:02:c9:01:07:fc:c7:12 > brd 00:ff:ff:ff:ff:12:40:1b:ff:ff:00:00:00:00:00:00:ff:ff:ff:ff > inet 192.168.0.2/24 scope global ib1
I was thinking of bonding rather than multihoming with the same address. Aren't there operational issues when configured as you show above ? -- Hal _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
