Reading these comments it looks like everyone runs into this all the time. As a counterpoint: never run into this on Azure, scaling up/down 20-30 vm's a day. Hope it stays that way...