Turning the Tables on VM Sprawl with VM Archiving and Resource Management

Posted 08/21/2013 by Commvault

Posted in

One of the biggest problems in today's virtual infrastructures is the waste of resources as a result of virtual machine (VM) sprawl. VM sprawl is an unintended side effect of the flexibility inherent in virtual infrastructure deployments. Before virtualization entered the mainstream, the complexity and cost of physical infrastructure deployment served as an automatic vetting process for applications. The relative simplicity of deploying virtualized resources has reduced these barriers, which sounds great on the surface, but many organizations now realize that this is a double-edged sword: we gain the flexibility and agility to enhance productivity by deploying resources more efficiently, but we also create a new management nightmare in the form of orphaned VMs.

Causes of the Sprawl

A great use case for virtualization is to spin up resources for short-term projects and then decommission and redeploy those resources when a project is completed. However, in reality, most of the VMs created for these use cases never get decommissioned and continue to consume valuable resources long after their useful life has ended. Test/Dev environments are notorious for this type of VM usage pattern.

Decommissioning usually doesn't happen for one of a couple of reasons – primarily depending upon who is responsible for the end-of-life process. If that person is the Systems/Virtualization Administrator, VM decommissioning is often left undone due to fear of the unknown – because it's hard to be sure if an orphaned VM is truly orphaned, Admins will usually leave these resources alone just to avoid a call from an angry end-user wondering why on earth their VM is gone. If the responsible party is the end-user – perhaps a developer or other engineer – they may avoid decommissioning immediately for fear that they may need the VM again soon. Then, when these resources aren't needed for some time, they're forgotten.

Killing the Zombies!

Several virtualization management tools have emerged to help admins identify these orphaned VMs (i.e., idle, stale, or zombie VMs), but the vast majority of virtual environments have not deployed advanced VM management tools due to time, cost, or some mixture of the two. It's also worth noting that while these virtualization management tools can help identify and take a few remediation steps to eliminate idle VMs, they are primarily reporting tools that show you where you should take action. That's one of the primary reasons I believe that 'Server Provisioning and Configuration Management' are in the 'Trough of Disillusionment' in Gartner's 'Hype Cycle for Virtualization 2013.' The tools available to handle these tasks still leave something to be desired in our world that desires more set-it-and-forget-it types of automation.

Gartner's Trough of Disillusionment

That's why Commvault's announcement of new Simpana VM Archiving functionality for VMware vSphere is game-changing! While most environments don't have a virtualization management tool, backup and recovery tools are ubiquitous. Because, as a backup and recovery tool, CommVault has visibility to all of the data under management, our software can see virtual machine disks (VMDKs) that have been sitting idle and kick off a customizable, policy-driven process to help eliminate VM sprawl - automatically!

Simpana 10 VM Archiving allows you to identify idle or stale VMs, power them down, use Storage vMotion to move VMs off to less-expensive storage – whether tape, disk or cloud – and then stub/archive these VMs off of primary storage, all while maintaining the ability to restore quickly in the event that an archived VM is needed. This process removes the risks associated with eliminating VM sprawl.

Risk Removal with VM Sprawl

You can archive your VMs…but you can get them back, too

Lots of vendors talk about 'VM archiving' in the sense of just vaulting VMDK data to a deep tier of storage (usually tape or other removable media). However, it's more appropriate to call these solutions 'VM vaulting' because the options for restoring vaulted VMs from this tier are no different than traditional full-VM restores from backup copies.

Alternatively, Commvault's more intelligent VM Archiving functionality places a stub before archiving the VMDK so that the archived VMs can be easily and quickly identified and recalled in a few mouse clicks using the Simpana 10 GUI or combine VM recovery with other processes in an automated workflow.

Simpana 10 GUI - VM Recall

Even better, Commvault's OnePass™ technology reduces the time and resources required for these activities by combining backup, archive and reporting into a single process. This eliminates the need to scan the file system multiple times while still extracting maximum value from these operations.

In addition, there are multiple recovery options for these VMs. Archived virtual machines can be recovered as complete virtual machines, VMDKs, files and folders. A virtual machine can be restored to the same ESX or a different ESX host; VMDKs can be restored to the datastore of an ESX.

VM Restore Options

And while all of this is great, the real secret sauce that sets Simpana VM Archiving apart from the rest is the fact that it is totally integrated into our industry-leading Simpana 10 data management platform and free to customers using our Virtual Server Agent. This brings unprecedented value for customers extending beyond data protection into full management and access capabilities including end-user access for restores, comprehensive reporting capabilities, and much more.