Server virtualization makes it possible to run multiple applications and operating systems on fewer hardware resources, which appeals to many IT managers looking to improve utilization. According to a recent Forrester Research poll, respondents have virtualized about one-quarter of their servers and plan to have close to 50% virtualized in two years. As enterprise IT teams look to broaden their server virtualization deployments, it’s important to get in front of the management challenges.

For those who are struggling with how to manage virtual machines, here are answers to six important questions.

What’s so tough about managing virtual servers?
Some will tell you that managing virtual machines varies little from managing physical servers, and others will say it depends on what you’re managing. But all agree you need to have a comprehensive management plan in place before widely deploying virtualization in production environments.

“Management is not a single discipline. It can range from business continuity planning to patch management,” says Andi Mann, a research director at Enterprise Management Associates. In the case of business continuity planning, virtual servers could be considered easier to manage than physical servers, Mann explains, but when it comes to patching multiple systems, the virtual world introduces complexities. “You can’t always be certain if all virtual systems are patched, and obviously that’s a problem,” Mann says.

Consistency and standardization also become a bigger issue when managing virtual servers alongside physical machines. The perks of virtualization include easy-to-deploy resources, and that demands IT managers have predefined configuration parameters for application and database servers, for instance. Experts say keeping configurations accurate and up-to-date becomes more critical in the virtual environment because configuration drift is more apt to happen on virtual machines. The same goes for patching.

“The focus shifts to managing templates and preventing drift,” says Jasmine Noel, principal analyst with Ptak, Noel and Associates. IT managers would ideally create a standard template that details the operating system, vendor software, patch levels, custom code and more. The template would be maintained so that every new virtual server deployed remained consistent with the predefined standard. Patching would also become part of the template, Noel says.

Beyond maintenance and availability management, another key management issue is performance. The complexity of a virtual environment makes determining the root cause of performance issues a more daunting task, industry watchers say.

“Performance management becomes trickier because for the more difficult problems you’ll need to understand how physical server issues manifest in the [virtual machines] and vice versa,” Noel says.

While virtualization provides flexible resources, multiple virtual machines residing on one box compete for the same resources, and IT managers need to keep that in mind.

How do I curb virtual server sprawl?
Virtualization offers ease of deployment, which can become a bit of a Catch-22 scenario for IT managers. The faster servers can be provisioned, the more it seems they are in demand – and that quickly leads to too many virtual machines.

IT managers and industry watchers say controlling virtual server sprawl requires the same processes and auditing that would be used in physical server deployments to ensure only as many machines as needed get provisioned.

“We have it set up so that no one has the rights to add virtual servers without requesting them through IT,” says Marc Kraus, manager of IT infrastructure at Merkle in Lanham, Md. “We run weekly scans as well to keep that in check.”

While policy-based management and inventory tools can help IT stay on top of the number of servers, IT has to be disciplined about putting processes in place to prevent virtual sprawl from corrupting the success of a deployment.

“People know we are able to bring up a new virtual server and turn that around quickly so the requests have increased. We basically have had to push back a bit against server creep,” says Albert Ganzon, director of network services and engineering at international law firm Pillsbury Winthrop Shaw Pittman LLP in San Francisco.

Industry watchers suggest adopting a server life-cycle management process in which a virtual or physical server’s purpose and status from creation to retirement is tracked. Failure to curb lax deployment habits can exacerbate other challenges around managing virtual machines, such as patch management.

Tags: Albert Ganzon, analyst, Andi, application, Associates, availability, bit, box, business, business continuity, Case, Catch, cause, check, code, complexity, comprehensive management, configuration, configuration parameters, Consistency, continuity, creation, creep, custom, database, database servers, demand, deployment, director, discipline, drift, ease, engineering, enterprise, enterprise management associates, environment, Failure, firm, focus, Forrester Research, front, Guide to Server, Guide to Server Management, hardware, hardware resources, industry, infrastructure, instance, inventory, issue, Jasmine Noel, Lanham, law, LLP, maintenance, management, manager, Mann, Marc Kraus, Md., Merkle, mind, multiple systems, network, Noel, number, one-quarter, operating, part, patch, Patching, performance, physical servers, Pillsbury, place, plan, planning, poll, poll respondents, problem, problem mann, process, production, production environments, Ptak, purpose, research, research director, research poll, retirement, root, San Francisco, scenario, Server, Server Management, server virtualization, software, sprawl, standard, standardization, status, success, system, task, template, top, utilization, vendor, vice, virtual environment, virtual machines, virtual servers, virtual systems, virtual world, virtualization, Winthrop Shaw Pittman, world