Customer has two nodes, both connected to two racks/shelves of disks (total 24 disks). One node has an aggregate using all 24 disks and the other node has another aggregate using all those same disks. Each aggregate has an SVM. Why not have one aggregate with two SVMs instead?