HA cluster

Peter Ross Peter.Ross at alumni.tu-berlin.de
Sat Jun 12 11:45:19 GMT 2004


On Wed, 9 Jun 2004, ROBERTO SIANO wrote:

>
>                        ( net )
>
>                           |
>                           |
>                           |
>                       ----------
>                      |  switch  |
>                       ----------
>                       /        \
>                      /          \
>              ----------      ---------- failover
>             |  active          backup
>                router  |    |  router  |
>              ----------      ----------
>        load-balance \          /
>                      \   VIP  /
>                      ----------
>                     |  switch  |
>                      ----------

This part can be solved using VRRP.


>                       /      \
>                      /        \
>               ----------     ----------
>              | server1  |   | server2  |
>               ----------     ----------
>                      real servers
>

This depends on the functionality, what kind of data the servers
use, whether there is a database involved etc.

BTW: Your scenario has two switches as a single point of failure..
Replicate them too if possible (You may use only two of them if you decide
two separate the "two ways" and combine before and after router switch by
using VLAN).

Regards
Peter


More information about the freebsd-cluster mailing list