Migrating from VE LTM to Physical 4000s with minimal downtime

We have a new pair of 4000s that we would like to migrate to, hopefully with minimal downtime.

 

My plan was to set up the new 4000s (management ip, admin access, licensing et al), and copy over the scf file so all the nodes, pools, vs etc are set up. I would not activate the VSs at this point.

 

The requirement is to 'test' the new setup first by activating our QA pool and testing against that. I can't see how we can do this if we swap the ip addresses of the new pair to reflect the old pair, as this would be a complete transition.

 

I was thinking of doing the following:

 

Bringing the QA VS down on the old server - call it QA_VS Bring it UP on the new server -

 

Do this for each pool until we are assured of full functionality.

 

Does this make sense? Or am I missing something ? I realize I will have 2 pairs of F5s up and running during this transition, but as only one of them would be accepting connections for a given VS at a time, it should work.

 

Advice from anyone who has done this would be GREATLY appreciated!