problem i see with this is that you will have to copy your main database every (timeperiod you decide) then shut down the secondary system and run a script on the database to change all the pollers and fix the type of polling. Then restart the secondary system once the script is done having inserted the new copy of the database.
This is a very intense operation and will require manual intervention each time you do it.
once you perfect the script it will be manual to the point that you must stop Solarwinds on the secondary system until you insert the new database.
one other thing you might need to think about is frequency of polling. If you are combining all the devices from 3 pollers to one you will probably have to drop back on the polling to 3 times the normal poll cycle on you main system.
This really seems like a lot of work when you could just stand the box up as a spare solarwinds server and if the main fails just cut over to it.
The stopping and starting of Orion should be the easy part. We will use a simple batch script. I think I will only need to copy the nodes table (but that is what I am looking into). In our BD this is only about 6.8 MB. That copy should be quick. About the polling cycles. The mirror system is only going to ping the nodes.
Fail over is not an option. We need an off site solution that will run if we loss the main site.