Code review comment for ~paul-meyer/cloud-init:fix-raw-ephemeral-disk

No problem and thanks, Scott.

BTW, we've also started noticing mkfs failures on 14.04, which we think is related to an async partition table reread triggered by 'sgdisk -p /dev/sdb' (similar to the issue outlined here: https://github.com/Azure/WALinuxAgent/pull/183).
It might be a good idea to insert 'blockdev --rereadpt' before attempting mkfs. I can make a merge proposal for that and switch some of these other methods to using blockdev as well?