the same hardware, the same firmware versions, the same management, and so
on in your testing environment. The closer it is to the production environment,
the more confidence you’ll have that a successful test process will result in
success in production.
I heard of a company that tested an implementation in the test environment and
then moved to production, and within a few days, started suffering blue screens
and huge service outages. It was caused by a mismatch of driver and firmware in
production that was not present in the test environment. There are many similar
stories, unfortunately.
This does not mean that the testing environment needs to be on the same scale as
the production environment. I may have 200 servers in production and 2 in
development, and that’s generally OK (unless you are performing scalability
testing). What is important is that the components are the same.This list is not definitive, and your organization may have other processes, such as
adding to change-control systems, but certainly these actions should be considered the
minimum actions to take.
Deploying Hyper-V Servers with SCVMM
You already have a method to deploy operating systems to your physical servers. This
could be over the network from a solution like System Center Configuration Manager
or Windows Deployment Services, installation locally from a USB device or CD, or
even some other process. That process may be well understood and tested and should
be maintained for your Hyper-V hosts as much as possible. Minimizing the number of
ways an action is performed reduces complexity and the chance of errors. If your
process is not optimal, there is a solution to deploy Hyper-V hosts (and even file
servers in the 2012 R2 release and above) included as part of SCVMM.
Rather than document the complete process to deploy a Hyper-V server with SCVMM,
I instead refer you to the Microsoft step-by-step detailed documentation available at
the following location:
http://technet.microsoft.com/en-us/library/gg610634.aspx
I do, however, want to cover the high-level process and specifically what a Hyper-V
host deployed using SCVMM will look like once deployed. SCVMM uses a feature,
boot-to-VHD, that allows a physical server (or desktop) to boot from a VHD file.
At a very high level, the following actions are performed:
1 . Windows Deployment Services is deployed to the same subnet that the servers are
being deployed to (or an IP helper-address is configured on the switch in the
subnet with the servers and points to the WDS server) and is configured inside
SCVMM as a PXE server (which allows SCVMM to configure and add required
images to WDS).