Many applications are now required to run 24X7 with little or no downtime. They are also expected to quickly recover in case of a natural disaster or internal outage. So how can you use Server 2008 to easily have highly available, scalable and manageable system even if an application is not designed to be cluster aware? One answer is to use a Server 2008 failover cluster along with Hyper-V Quick Migration.
With quick migration, you run server workloads and applications on virtual machines that are hosted and moved between clustered physical (host) servers. This enables you to provide a wide variety of services through a small number of physical servers and, at the same time, maintain the availability of the services you provide.
Quick migration combines Windows Server 2008 Hyper-V hypervisor-based virtualization with failover clustering in Windows Server 2008 Enterprise and Windows Server 2008 Datacenter. In quick migration, you run server workloads on virtual machines—guest operating systems running in turn on physical hosts. Each host server is then configured as a node in a failover cluster. If the server initially hosting the virtualized workload fails or requires scheduled maintenance, another server in the cluster hosts the virtualized workload.
If you think about this it is pretty cool. Each of the guest operating systems behaves like a cluster-aware application, with the cluster service providing health monitoring and automatic recovery for the guest operating system.
Now Hyper-V Quick Migration is not live migration. For planed downtime a quick migration saves the state of the running guest virtual machine to storage and then moves the storage connectivity from one physical server to another. It then restores the guest virtual machine onto the second server. So how long does this take? The answer is it depends on the speed and sophistication of the storage and network hardware in the cluster. The other factor is the amount of information in RAM (state of guest OS) that must be stored before the move can be completed. For unplanned downtime the guest OS is not able to commit memory (state) to storage. So in this case the guest virtual machine is started from previously stored images. If you are using a backup program that takes snapshots of the running virtual machine you can recover to the most recent snapshot of the virtual machine. Here are some scenarios where Hyper-V Quick Migration may be useful:
· Scheduled maintenance for host (either hardware of software updates)
· For operating systems or applications that could not previously have been run in a cluster
· Maintaining the availability of servers consolidated on to virtual machines.
For more information on this topic please see the following links:
http://download.microsoft.com/download/3/B/5/3B51A025-7522-4686-AA16-8AE2E536034D/Quick%20Migration%20with%20Hyper-V.doc
http://blogs.technet.com/roblarson/archive/2007/12/17/building-a-host-cluster-with-hyper-v-beta-1.aspx
http://download.microsoft.com/download/3/B/5/3B51A025-7522-4686-AA16-8AE2E536034D/Microsoft%20High%20Availability%20Strategy%20White%20Paper.doc
Until next time, enjoy!
