Re: Failover

[Derick Siddoway <derick@panther.bitflood.net>]

On Wed, Apr 30, 2003 at 11:53:15AM -0400, Roy Morris wrote:
> The big problem with any of these fail over methods is that they 
> only understand a 'down' item and in most cases the item, be it 
> an adapter or vip is actually alive well the routing daemon or 
> application is dead. Checkpoint on Nokia was famous for having 
> these problems, while vrrp was able to reach the interface and 
> therefore it was alive, the checkpoint application its self was 
> dead (nothing new for Checkpoint there!) and no action taken.

Right, so you don't just ping the active host, you test the
application on that host.  If you're doing failover on a
database, for instance, you actually execute some SQL for
the "aliveness" test.  At least, that's how it worked on both
sun and veritas clusters.

-- 
Derick Siddoway      Yay, sleep!  That's where I'm a viking!
derick@bitflood.net            - Ralph Wiggam