Windows Server 2012 features in-depth

In the final part of our virtualisation series, we dig deeper into some of the new features of Windows Server 2012

Ben Regis
11 Jan 2013

So far in this series we’ve discussed the benefits of virtualisation and walked through the stages required to implement virtual machines. In this final part we’ll dig deeper into some of the excellent new and improved features in Windows Server 2012.

Live Migrations

Once you move beyond a single host server the ability to move virtual servers between hosts becomes important. Live Migrations were introduced in Windows Server 2008 R2 but have been greatly improved in Windows Server 2012, and you can now move running VMs between servers even if they don’t have any shared storage.

Configuring Live Migration is simple. Both servers need to be members of a domain and you need to configure a few settings for each server, such as which authentication protocol to use. When you initiate a move, Live Migration will copy the configuration information to the destination server. Virtual hard disks (unless shared storage is used) and the VM’s memory state will also be copied, and any changes will be synchronised.

Hyper-V: Live Migration

Once the source and destination are almost identical, the VM will pause for a fraction of a second. Any remaining changes will synchronise, and the VM will stop running on the source server and start running on the destination server. On a fast network the final changeover can happen so quickly that users will be unaware that anything has changed.

Another option is to perform a storage migration. You could use this if you had a Hyper-V server with VMs running on internal hard disks but you then added a SAN to the network. In this case you’d perform a Live Migration as above, but select the option to just migrate the storage. This would leave the VM running on the server while the files are moved to the SAN, again without any noticeable downtime.

Teaming, clustering and SMB

As we’ve already discussed in this series, it’s important to remove single points of failure by using fault-tolerant components. Windows Server 2012 can help in this area with simplified implementations of network adapter teaming and clustering.

By configuring two adapters into a network team, if one of the adapters fails the server will still be able to maintain its network connection. You can also use teaming to increase bandwidth. Windows Server 2012 handles the teaming so you don’t have to use identical adapters, although it’s best to team adapters that run at the same network speed.

You’ll have some of the functionality normally associated with a SAN without the associated high costs

You can make all adapters active or set one as a standby that will only be used if another adapter fails. The simplest teaming mode is “Switch Independent” and as the name suggests, it will work with any switches. Note that in certain situations, such as heavy inbound and outbound traffic, you may need to select a different teaming mode and configure your switches to match this.

Windows Server 2012 includes version 3 of the Server Message Block (SMB) protocol. While initially this may not appear to be the most interesting of updates, there are actually a large number of benefits. In terms of Hyper-V this allows VMs to run from a file share on Windows Server 2012 file servers, providing a greater number of virtualisation infrastructure scenarios.

For example, you could use shared direct attached storage between two file servers, create file shares using this storage and then store all the VMs on the shares. With improvements in clustering in Windows Server 2012 it’s relatively simple to cluster the two files servers removing a single point of failure. Combine this with Storage Spaces, which provides flexible use of disk storage, and you’ll have some of the functionality normally associated with a SAN without the associated high costs.

Read more about: