We had such a bad experience upgrading from 6.3.2 to 6.4 that we had to do an emergency rollback. Hold off on the upgrade if you have a lot of virtual HBA ports on your switch fabric. There are still issues with the new NPV feature and how those ports are represented and detected. I've been working with internal NetApp people for awhile to get these issues addressed and corrected.
Other than that, the upgrade itself should go just fine if you follow Jim's suggestions.
In theory, the upgrade should just work. In practice, there is quite a variety of different things that can go wrong, and it is helpful to have several potential fallback plans.
NetApp Support does accept proactive upgrade cases. It might be good to open a case informing Support that you will be performing an upgrade, and when. Then if there is any issue, an OCI expert in the Support Center will already know your serial numbers, versions, site locations, and all the good details. This can speed resolution of any issues. You may be able to talk them into a walk-through as well, if you like.
If you're running on VMs and can snapshot them, do so prior to upgrading. It provides a straightforward, reliable, and time-bounded last-ditch rollback option.
Make a copy of the entire sanscreen\cognos\c10-64\deploy directory, in case you need it to recover your reports.
Have valid web interface backups of each OCI instance and the DWH. If you can't generate these, don't proceed.
That's what's coming to mind for me at the moment.