I have always wanted to make a ceph cluster with one disk per node, just to see how it would behave, from a performance view point it would suck, but from a fau

I have always wanted to make a ceph cluster with one disk per node, just to see how it would behave, from a performance view point it would suck, but from a fault tolerant viewpoint it would be amazing.