hmm...
It works really good..
.. as long as you
- don't install a changed policy
- don't have a fail-over
Well, this may be just our specific problems... (CP first stated that we're the only, who have these problems... after a loooooong time they mentioned that another customer has similar issues.. CP is now investigating this case with an external routing specialist... I will post any updates to this issue here [1])
If you still want to try it, I suggest you
strongly to try this first in a isolated test-environment!
Try to use the same devices (also the same IOS-Versions in case of cisco-devices).
Test fail-overs and policy-installations when one node isn't running.
Check if you still have all routes on the peripherial ospf-devices after a fail-over (we have seen some really weird stuff!)
Check if your other OSPF-devices are capable to do a graceful restart (RFC3623 (section 4)) which should be configured to ensure a minimal interrupt.
[1]
SPLAT PRO, OSPF and Cluster => Interrupt while Policy inst.