Operating system upgrades are unavoidably disruptive, so it's important to perform all of the preparations needed...
to mitigate interruption and ensure a smooth rollback path in the event of problems.
Preparation starts with upgrade familiarity. No matter how many upgrades you've been through, it's always a good idea to test the upgrades and rollbacks in the lab first. This should include testing the host system, as well as any virtual machine (VM) upgrades (including any resource provisioning changes to accommodate the new OS). This kind of testing can re-familiarize staff with the procedures and find workarounds for any possible glitches well in advance of the actual production upgrades.
Plan workload balancing to redistribute VMs while the host server is upgraded, and understand the order needed for upgrades. For example, it is probably more important to upgrade a domain controller or DHCP server before upgrading other applications. Some organizations will deploy upgrades in phases and allow ample time for shakedown testing and optimizations before moving on to the subsequent phase.
Also prepare for upgrades by performing full backups of each server and VM. The actual backup tool or protocol can vary dramatically between organizations, but backups should generally include all boot, partition and system data needed to operate the system. If the upgrade fails, restoring the backup may be the only way to recover the server and resume normal operation.
Benchmarking performance before and after the upgrade process can also help to ensure objective performance levels and offer guidance for configuration changes or other optimizations that will allow applications to get the most benefit from Windows Server 2012 R2.
Distilling the installation requirements for Windows Server 2012 R2
Find out the upgrade limitations for the latest version of Windows Server
Make sense of Windows Server licenses