[Deploying Sakai] Dataguard failover mechanisms

[Todd Williams]

I'm beginning to work on an initial Dataguard config for our early 
dev/proof-of-concept Sakai implementation.  I'm doing this with 11gR2 on 
a primary and one physical standby.  The big question I keep running 
into is how to handle switchover/failover scenarios.

My understanding is that Oracle's thin jdbc driver is capable of 
receiving the various fast-application notification signals, but that 
the actual failover is dependent upon the application *doing* something 
with those notifications.  So I guess this leads me to ask if anybody 
been able to do this with Sakai 2.6?

In an earlier thread Adi at Michigan mentioned that they were never able 
to pin down TAF for RAC, so I wonder if there are similar outcomes with 
FAN and DataGuard.

So generally I'm interested in how folks are handling Dataguard 
fail/switchover.  I suppose there's the option of whole-cluster failover 
via network load-balancer; however, I'd love to retain the flexibility 
of failing the DB tier w/out having to fail the entire application cluster.

Thanks in advance for your input!

-- 
Todd Williams
tcwill at ufl.edu
CNS Open Systems
University of Florida