Get startedBring yourself up to speed with our introductory content.
Microsoft Nano Server overhaul a disappointment to some
Microsoft's course reversal on Nano Server in Windows Server 2016 will make some Windows administrators hesitant to come along for the ride in the next server-related innovation.
As recently as late 2016, Microsoft hailed Nano Server as the heir apparent to Server Core.
Microsoft Nano Server's inclusion in Windows Server 2016 excited experts, who buzzed about Nano Server's smaller footprint of 400 MB -- quite a reduction from a Server Core install of about 6 GB. Microsoft offered this new install option for specific services in the data center, such as Hyper-V clusters and scale-out file servers. The company said the smaller footprint -- mainly achieved by pulling the GUI -- would prevent attacks and reduce the number of patches.
However, Microsoft Nano Server lost those touted capabilities with the 1709 release of Windows Server 2016. Microsoft stripped Nano Server's functionality even further to make it available only as a base image for containers and so that it runs in a container in a container host. As Nano Server shrinks to around 80 MB, Microsoft removed infrastructure features, such as Hyper-V. Nano Server does not include Windows PowerShell, .NET Core or Windows Management Instrumentation by default. It lost the servicing stack to add roles, features and updates in the OS. To patch or update, the container redeploys via Docker, which also handles troubleshooting.
This change represents a failure for Microsoft and its customers.
Early adopters got burned
Microsoft's message is: Don't believe what we said about Nano Server -- sorry if that's why you upgraded to Windows Server 2016. Now that Microsoft Nano Server is for containers only, anyone who uses it as a file server host or a Hyper-V host must migrate before support ends in spring 2018.
Microsoft's message is: Don't believe what we said about Nano Server -- sorry if that's why you upgraded to Windows Server 2016.
When Microsoft scraps a major feature it promoted during Windows Server 2016's launch cycle, it means the company thinks the move won't affect manycustomers. Most IT pros I talk to who have Windows Server 2016 licenses choose to deploy Windows Server 2012 R2 for now. Proceed with caution before you deploy workloads that use Nano Server or Server Core, because of the murky situation and unclear future for each streamlined OS.
Server Core just isn't worth it
I don't particularly care for Server Core as an installation option. It causes administrative and day-to-day headaches that far outweigh perceived benefits, such as reduced attack surface and less patching. The administrative tools aren't great -- try to configure firewall ports properly for remote procedure call and the remote server administration toolkit on the first try -- although PowerShell is sufficient now that it's matured. Plus, Server Core doesn't save money.
I liked Microsoft Nano Server because it stripped down and refactored Windows Server; it represented Windows for the future. In contrast, Server Core feels like flying with blinders on. Unfortunately, Microsoft did not significantly innovate with Server Core in Windows Server 2016. I recommend Server Core for only sensitive deployments and prefer the convenience of the full GUI for almost every other situation, even if I primarily administer Windows Server with PowerShell.
Containers are overhyped
It's an odd choice to make Nano Server only available for containers. Midmarket and Fortune 500 companies with thousands of Windows Server licenses will not immediately go all-in on containers. Not every organization can buy into the DevOps continuous integration mindset. I have no doubt that container-like technology will prevail. But Windows Server is a product with broad appeal. Microsoft, rather than make a better non-Server Core infrastructure product, spent its time courting everyone's Docker obsession.
The original Microsoft Nano Server represented a move forward for the company. The new version is a step in the wrong direction.
Join the conversation
1 comment