If you're testing the Technical Preview of the next version of Windows Server (or Windows Server 10, for lack of a better name), it helps to understand the areas where improvements and new features are available over Windows Server 2012 R2. In this series, I've covered some important topics already, and I'll be picking through the rest over the next few weeks. Remember that many more new features will be added and enhanced over the next many months until public release, but I'll dig into those when available.
In this article, I'll cover the new Failover Clustering features included in the Technical Preview.
One major enhancement to Failover Clustering, called Cluster Operating System Rolling Upgrade, is the capability of upgrading existing Windows Server 2012 R2 cluster nodes to the Technical Preview without shutting down the Hyper-V or Scale-Out File Server workloads. Obviously, this ability is intended to eliminate downtime during upgrades. What happens is that during the upgrade, the cluster will continue to run as a Windows Server 2012 R2 node until each node in the cluster has been fully upgraded to the Technical Preview.
This is a huge gain, because in the past the entire cluster would have to be taken offline during the installation of a new operating system on each node. Imagine being able to upgrade the entire backend infrastructure without forcing users to logoff and wait until the process is complete. Also imagine performing upgrades during production hours instead of having to wait until after hours or the weekend.
Additionally, this feature eliminates the need for creating a new cluster just for upgrades, which also removes the need for additional hardware.
Keep in mind that after a cluster has been upgraded, it can no longer be downgraded, meaning don't install the Technical Preview in a production environment! This feature also only works for Windows Server 2012 R2 to Technical Preview upgrades – it does not work with earlier versions of Windows Server and Microsoft has stated it has no plans to back port this feature.
In addition, as part of the overall Failover Clustering component, Storage Replica services (covered already) will add further value to clustering by delivering server-to-server and cluster-to-cluster replication, and eventually, server-to-Cloud synchronization and mirroring.