Once you surmounted the toughest task — getting a “yes” from all the people involved to start slowly migrating your physical server environments to virtual environment — you need to know the common mistakes made by most people going from a physical to virtual environment, and this three part series of posts will help you do just that.
Mistake #1: Thinking that the Job is Done After the Servers are Virtualized
The first and the biggest mistake folks make is they don’t realize that your first task is virtualizing servers, network components like Network Interface Cards (NIC), switches, storage, and much more. Remember now you need to coordinate folks from different departments and communicate. You need to communicate your intentions to the Network folks so your virtual switches can communicate with physical switches, for example. There are so many things that could go wrong, so communication is the key.
Mistake #2: Counting on Cost Savings Before They’re Actually Realized
People assume they’ll save a lot of money when they go to a virtual environment. I’m not saying it’s not true; for instance, it could save you money by consolidating multiple underutilized servers into fewer servers that host multiple virtual machines (VM) and give you more use of the high-end processing power today. But remember you can’t just pick some cheap high-power servers. Many virtualization products have a hardware compatibility list (HCL) that tells you the approved hardware for the appropriate virtualization products.
Mistake #3: Not Getting Proper Training Before Installing and Configuring Your Virtualization Environment
So, now you have the coolest, meanest, and most expensive hardware approved by your management and on the HCL. You can’t wait to get started, but let me ask you: How many people in your department are trained to install, configure, and manage your virtualization environment? Many products might be easy to install, maybe a few Next buttons on some wizard, but at the end, you have to configure it fully optimized for your environment. The more people you have trained, the better it would be.
Mistake #4: Poor Load Planning
Have you looked at the load on your current servers? Do you know what the CPU and memory utilizations are? Maybe you need super-fast servers for your specific requirements, like your databases, but what about all those file and print servers? Maybe you are at 2-5 % utilization most of the time. “How many virtual Servers to run on these expensive VMware ESXi hosts?” is a good question to think about while planning for virtualization. I might consolidate a few File and Print Servers on one big powerful host running ESXi.
Mistake #5: Assuming Virtualized Servers Take Care of Themselves
In the past you had physical servers, and people took care of their own servers. Who’ll take ownership of these virtual servers now? Who looks at the logs on the host and on the VMs? You should have a procedure for when to look at the logs. I hope its every 12 or 24 hours. It might be too late by the time there’s a problem, so appoint someone to look at the logs regularly.
This post is excerpted and reprinted with permission from 14 Common Virtualization Mistakes — And How To Avoid Them by Dheeraj (Raj) Tolani