Mainstream support for Windows Server 2012 R2, the server version of Windows 8, is ending in just over a year.

That means clients with large Windows clusters now have to start thinking about replacing or upgrading these to Windows Server 2016. The good news, however, is the upgrading process can now be much less time-consuming and expensive, plus demand vastly reduced downtime, compared to procedures that had to be adopted in the past.

A cluster is a group of independent systems, called nodes, working together as a unified resource, providing a single name for clients to use and a sole administrative interface. This arrangement guarantees data is consistent across nodes.

Windows Server 2016 is an operating system, part of the Windows New Technology (NT) family, developed alongside Windows 10.

With previous versions of Windows Server, the upgrading procedure normally involved building a separate set of servers and allocating new storage logical unit numbers (LUNS). These are unique identifiers for designating physical or virtual storage devices executing input or output demands with a host computer.

Once the new cluster was up-and-running, that upgrade process also formerly involved migrating roles, via further new installs and backups or restores. In addition, in the case of Hypervisor (Hyper-V), the virtualisation platform in the Windows Server providing the software allowing virtual machines (VMs) to run, it included importing and exporting these.

Now, with Server 2016, comes the ability to have a Mixed-OS Mode Failover Cluster. This allows different versions of a node to be grouped together in a fault-tolerant arrangement. It enables the 2012 R2 and 2016 servers to run together in a single cluster, enabling each node to be rebuilt or replaced individually, and then permits the whole cluster to be upgraded to 2016 functional mode.

This improvement delivers the very significant benefit that upgrades of entire Hyper-V clusters can now be carried-out with zero-downtime. This is provided the constituent nodes have central processing units that support Second-Level Address Translation (SLAT) – Hyper-V technology within Intel and AMD processors. It also assumes there is sufficient capacity in the cluster to maintain high availability when each node is removed for rebuilding.

Put simply, the process of upgrading each node consists of:

  • Move (drain) any running VMs or roles onto another node
  • Evict the node from the cluster
  • Rebuild the node as Server 2016 (Microsoft discourages upgrading the operating system)
  • Restore the node to the cluster
  • Move the required VMs and roles back to the rebuilt node

Once all the nodes have been rebuilt, you can raise the cluster’s functional level to 2016.

This new feature is sure to save businesses considerable time and money, as it removes not only the requirements for extra servers and enterprise storage space but also, in many cases, overtime payments – as IT teams often had to migrate the VMs and roles out-of-hours.

For more information on upgrading Windows clusters or Windows Server 2016, please contact Opal Wave.