The ability to quickly and easily create VMs allows environments to scale on demand and is a key part of why virtualization gives businesses such an advantage. But this important feature can also cause a serious problem. When an employee can provision and deploy a virtual machine with nothing more than a few mouse clicks, there is a dangerous potential for VM sprawl -- the term that applies when the number of VMs in an infrastructure reaches a point where administrators can no longer effectively manage and maintain them. Left unchecked, these VMs -- which may or may not be needed -- consume valuable computing, memory or storage resources.
VM sprawl, as any virtualization admin knows, can cripple an otherwise healthy environment. In fact, the reason this sprawl is so problematic is that the underlying cause often stays hidden until it manifests in resource shortages. For this reason, administrators need to be proactive with VM sprawl control strategies.
Here are five quick tips that will help you stay on top of VM sprawl in your environment.
Develop policies to prevent sprawl
Strict business policies are one way to prevent VM sprawl. When employees need to justify a specific reason to create a VM, they are more likely to think twice before requesting resources they may not need. This type of restriction may not be practical for all businesses, but assigning users permissions can limit VM creation. At the very least, employees should be familiar with a set of guidelines or best practices for creating new VMs. Businesses should also have policies in place for determining how many VMs a server can host.
Use automated reporting to stay on top of resource use
Of course, employees don't always follow best practices, so you should also consider software that can report resource use and virtual machine capacity as a part of your VM sprawl control strategy. Many capacity planning tools can alert admins if resource usage reaches a certain threshold, warning you of a potential VM sprawl problem before it gets out of hand. Even built-in tools, like VMware's esxtop utility, can generate resource usage reports.
Consider lifecycle management tools
One of the most common causes of VM sprawl is VMs that continue to run after they're no longer needed. A lifecycle plan can help prevent unneeded or outdated VMs from consuming resources. Many VMs are needed for only a short time. Lifecycle management tools can help you keep track of which VMs are running. If possible, create VMs with time-based policies so they're automatically removed once they are no longer needed.
Create a VM library
Creating a library of standardized VM templates can be an important part of VM sprawl control. When users create VMs from one of these base images, IT departments can easily identify the purpose of a VM and, if appropriate, know when it can be deleted. Creating consistent VMs with built-in limitations also prevents one VM from consuming too many resources. Keeping track of software updates or patches is much easier when you have similar VMs.
Try VM archiving
Even idle VMs consume resources, so part of your VM sprawl control strategy should include removing those idle virtual machines. But not every idle VM is expendable, and you may not want to simply delete them forever. By archiving VMs that you may need later, you can reclaim memory and computing resources that otherwise would be wasted. These archived disk images, usually stored on inexpensive storage media, can be pushed back into production without much trouble when they're needed.