HW renewal: Add 2 nodes to RAC cluster and shut down old ones: procedure ?

Hi,

We have a 2-node RAC database running over here in 11.2, on, say, dbnod1 and dbnod2.

These two servers are old hardware, which is going to be replaced, but the underlying SAN

of our ASM diskgroups *won't be touched*. So there's no physical move of data.

D'you think it's feasible to set up 2 DB servers made on this new hardware (dbnod3 and dbnod4),

start instances up on both, and shut down dbnod1 and 2 ? Pb is Grid infra I guess: if you add

nodes to a cluster, how easy is it to shut down original machines and have new ones now up