Basically you have two choices:

The cluster approach: None of the services are running on one node, until the other one crashed. Then the second node brings up proxy,
secondary IP addresses.

The two-proxy approach: Both proxies are running, but one is set to completely different IP addresses than the other. To switch over, you

have to change addresses around, in a sequence that avoids duplicating

an address on each server. Can be done, but generally a checklist is

recommended for the switchover. In this method, you switch over more

slowly, but you can use the second node for testing while production traffic goes through the first one.

In both cases, you tend to want all the BMgr services using secondary

addresses (public and private).

Craig Johnson
Novell Support Connection SysOp
*** For a current patch list, tips, handy files, books and
consulting services for BorderManager, go to
http://nscsysop.hypermart.net ***