We made a large storage purchase and NetApp "threw in" three "free" FAS2220 12-disk pairs. With 12 disks for a pair of filers, they're basically next to useless. By the time you take out two disks for spares, 4 disks for dp parity, you are left with 6 data disks (7 if you do raid-4 on one head like above). We thought long and hard about how to distribute the disks, and we settled on evenly distributing them between the two controllers, so as to be able to take advantage of the processing power of each filer. 6 disks on each filer: 1 spare, 5 disk raid-dp aggr0. We shrunk vol0 as much as we could to free up more space.
As much as we hate wasting disks with raid-dp, it makes upgrading firmware non-disruptive and that's why we use raid-dp everywhere.