You'll have to forgive me because I can't put my hands on the document whose recommendations we followed, but we followed them in the way that was outlined above and were successful. But aside from just following a document blindly, I can see a number of potential pitfalls if you don't properly license the cluster before you restore from backup. For one thing I would assume you wouldn't be able to do the restore correctly.
I know this whole thing is icky and I understand you not wanting to operate without a net during that window when you essentially have a blank cluster. However, I don't think there's any other way to do it. Just MAKE SURE you have the media, the licenses, the documentation, and good backups. Remember, too, that you always have the old hard drives that you could pop back in if there is a tradegy.
To answer your other question, after the hardware was switched out and we began the reinstall, we gave the servers the same name and IP that they had originally before we did the restore.
One other thing worth pointing out is that the above steps just got us back to where we had begun except we had new hardware in the servers. We were running the same version of everything and with the same configurations as restored from backup. The step AFTER THAT was to do the upgrade to the new version.
After having gone through it, my suggestion is to give yourself a couple of windows in which to complete the entire operation. We actually did it over two weekends. It was a little less nerve-wracking that way. On the first weekend we performed the above tasks. On the second weekend we did the upgrade. This allowed us to monitor the cluster with the new hardware for a few days and feel good about its stability before we went to 8.6.
