The code that "wipes" a disk is wildy inefficient and we also found it's prone to race conditions. ()

Issues I've found.
1. we're opening the underlying disk multiple times for a single "wipe" operation
2. we're dd'ing the first 32MB and last 32MB of the disk as well as gpart destroy/gpart create and another gpart destroy (The gparts are not needed since the dd is plenty sufficient enough to do it)
3. we're subprocessing out to run these commands multiple times in a single method