TimRead wrote:We brainstormed much the same ideas as you did in coming up with a script which would
Are you sure that the complexity and risk this adds is worth the performance gain you get from the extra SGA size? Remember that you'd need to ensure that the SGA sizes were increased back to the full size when the second DB was moved back to its original node. Achieving that reliably is also quite hard when you start to consider all the failure scenarios.
If you really are sure you want to do this, then you'd probably be best writing a Generic Data Service (GDS) resource type that allowed a start and stop script to be created that changed these parameters and performed the second service restart. The DB resources would then need a dependency set against these resources. Having said that, I would not go down this route as I judge the risks and complexity are too high and you'd end up lowering your availability.