“What risks do I run by using a vm for my vCenter server?” This question is one of the most frequently asked of me in the field. I like to encourage people to ask the converse question when planning for their virtual deployments: ”What are the risks of using a physical server for vCenter?”

A common answer I have been given by several clients is that they do not want the management server within the environment that it is managing. This is a sensible argument to me; it is a debate I have engaged in many times as an engineer and also as a consultant. There was a time when I would have opted for a physical server. Yet, when I started asking both of the above questions, the beauty and safety of a virtual vCenter server became evident. 

Much of the aversion to a virtual vCenter server may simply be based on comfort and knowledge levels of the ESX/ESXi design and functionality. One of the big concerns clients have is this: If the host within the vCenter server dies, users lose the ability to manage the environment. This is not necessarily true though. Below I have shared the top three reasons I like a virtual vCenter server:

1.   The whole concept of virtualization = shrinking the physical footprint.
2.   Ability to take easily manageable snapshots of your vCenter server.
3.   vCenter server is a virtual machine protected by HA (High Availability).

Did you catch the last one? It’s true!

HA will protect the vCenter vm. Your vCenter server is required for the initial setup of HA; once it is configured the hosts have their orders and will carry them out without hesitation. So if the host with the vCenter server vm fails, your vCenter server will be powered up on an existing host as will all the other vm’s on the failed host. In the interim, users still have vSphere client access to their surviving hosts.

With a physical server, many different hardware components can fail and render the server out of commission until a physical repair or part replacement can be performed. Even with the highest level vendor support, a part replacement can take over 4 hours to complete. One way to protect against this is using MSCS (MicroSoft Cluster Service) to create a fault tolerant physical configuration. However, this adds to your physical footprint and may increase the complexity of your environment.

There are many more pros and cons to consider when planning your virtual infrastructure, and I would love to hear some insights from people who are using both physical and virtual vCenter servers.

My thought is this: Plan implementations carefully, always weighing all sides before committing. If a physical vCenter works for now, we can always P2V it down the road!

Photo Credit: Snap-shooter