Slow news day at ComputerWeekly.com — at least based on the reruns they are playing of the Menzies Aviation Hyper-V case study. One thing that really stands out in that article is the apparent lack of understanding on the part of the IT staff at Menzies when it comes to high availability and shared storage:
One Hyper-V 3.0 feature is called Shared-nothing live migration. It allows administrators to migrate a virtual machine (VM) from one physical server with direct attached storage to another physical server with direct attached server.
It requires no clustering or shared storage between both servers as just a small connection such as a Gigabit Ethernet (GbE) cable will suffice resulting in zero down time.
For the cargo company that operates in 180 airports around the world, security and zero downtime were key criteria.
Both Hyper-V and vSphere allow administrators to place virtual disks on local storage or on a shared SAN. VMs using local disks are essentially pinned to one physical host and cannot be migrated to accommodate events like shifting resource requirements or host maintenance. Enter “shared-nothing live migration” — an interesting new enhancement found in the most recent releases from both Microsoft and VMware which allows administrators to move a VM and storage in a single operation.
Just the thing for environments that require zero downtime, right? Not exactly.
Important workloads with high uptime requirements need to be protected from physical host failures. HA is responsible for restarting failed VMs on surviving hosts, but this is only effective when VM disks are still available — on shared storage. Virtual disks residing on local storage will remain offline until the physical host is brought back online.
It’s that simple. Being able to migrate a VM and storage together is a feature with interesting use cases — none of which have to do with application availability and local storage.
For those interested in all the gory details, check out this excellent technical paper on VMware vSphere 5.1 vMotion.