Our company runs a hybrid Azure and on-premises environment, with most primary services hosted in Azure. On-site, we currently have a couple of aging ESXi hosts running four virtual machines: a domain controller, print and licensing services, a file server, and a phone-system server. Everything except the phone system runs Windows Server. As the hardware needs replacement, I'm considering moving away from ESXi and buying one small server for a domain controller plus a more capable server that would combine file, print, and licensing services while hosting the phone system as a Hyper-V virtual machine. Would this be a reasonable design, or are there important availability, compatibility, or recovery concerns I should consider?
4 Answers
Hyper-V on new hardware could work well, especially if the workloads are fairly light. Before consolidating everything, verify the phone system’s supported hypervisors, CPU and storage requirements, and backup or snapshot guidance. Also size the server for peak usage rather than just current averages, and make sure you have tested backups and a recovery plan.
A small Hyper-V host with separate VMs is usually preferable to putting every service directly on one Windows installation. Keep file, print, licensing, and phone services isolated where practical, use redundant storage and power, and maintain offline or otherwise independent backups. If the budget allows, two modest hosts would provide a much stronger recovery and maintenance story than one larger box.
You could also evaluate Azure Local if you want to keep the workloads integrated with Azure, but it is much more prescriptive than simply installing Hyper-V. The hardware generally has to be on Microsoft’s supported list, so existing equipment may not qualify and the licensing and management costs should be compared with a conventional standalone Hyper-V server.
I’d be cautious about running only one domain controller. Even in a hybrid setup, losing the sole on-premises DC can affect local authentication, DNS, and access to file shares while Azure services may still be available. A second DC—possibly as another VM or hosted in Azure, depending on your connectivity and identity design—would give you much better resilience.
The important part is planning for the failure of the physical host as well. Two DCs on the same server protect against a VM failure, but not against a hardware, storage, or power outage.