On Thu, Mar 8, 2012 at 12:07, Marco Shaw <marco.s...@gmail.com> wrote:
> I'm possibly looking at testing bonding in a "single network fabric"
> environment.
> I was thinking of giving NIC1 IP1, NIC2 IP2, and then creating BOND0 with IP3.
> Is this do-able?  All of the IPs are going to be routed through the
> same networking device.  I'm planning on an active-passive setup, so
> when I pull the network cable on NIC1, I'd expect IP3 to be service
> from NIC2.

You're confusing:
- Frame (Layer-2 -- e.g., Ethernet) high availability / load balancing, with
- Network (Layer-3 -- e.g., IP) high availability / load balancing.

Bonding handles the former, allowing more than one physical interface
to service frames (such as Ethernet).  There are different bonding
modes, including not sharing MAC addresses, failing over with the same
MAC address, sharing MAC addresses, etc..., using both layer 2 (e.g.,
IEEE802.3ad Link Aggregation, aka mode 4) and higher layer 3+ (e.g.,
ARP, which handles L2 MAC to L3 IP mapping, several approaches used in
several modes), and other facilities to update layer-2 (e.g.,
Ethernet) switch tables with the appropriate addresses

Bonding (usually Ethernet) options are somewhat akin to the various
options for Network (usually IP) high availability / load balancing
including not sharing the same IP address (e.g., Round Robin DNS),
failing over with the same IP address (e.g., active-passive/RHCS),
sharing IP address (e.g., Piranha/HA), etc...

It's very important to understand how the two differ, as well as how
they can both be used to reduce a single point of failure at various
levels.

E.g., use bonding to allow redundant network cards in a server, which
have redundant links, possibly to redundant network switches, use IP
solutions to allow redundant servers, etc...


--
Bryan J Smith - Professional, Technical Annoyance
http://www.linkedin.com/in/bjsmith

_______________________________________________
rhelv5-list mailing list
rhelv5-list@redhat.com
https://www.redhat.com/mailman/listinfo/rhelv5-list

Reply via email to