What is VM Sprawl?

Virtual machine (VM) sprawl is the uncontrolled growth and proliferation of VMs in a virtual environment. When this happens, the network reaches the point where you can't manage the VMs effectively and they start demanding excessive host resources. Typically, VM sprawl occurs as the result of the creation of virtual machines that stay unused or forgotten but still consuming host resources. You don’t realize it immediately after virtualizing your servers. VM sprawl happens over time and if it remains unchecked, say, at the end of one year, you’ll be surprised to see the number of VMs associated with the same number of hosts you created initially.

 

 

This is not good because VM sprawl can waste resources and contribute to resource contention on actively used VMs. It doesn’t end here. You’ll need to feed your new VMs more resources and you will likely end up looking for more procurement budget – which is definitely NOT what you want!

 

  

 

How to Identify VM Sprawl?

VM sprawl has no clearly defined symptoms, but there are some typical characteristics that can help you identify it.

  

 

#1 Look for Idle/Stale VMs

Idle or stale VMs are those VMs that are existing in your virtual environment consuming host CPU, memory, disk, and network resources but not actively used.

Stale VMs can also show other indicators such as:

  • No recent user logins
  • Files not modified for a long time
  • Not recently powered on
  • Orphaned and unused files

Also, look for the VM creation date to ascertain how long the VM has been existing and how little it has been actively used.

 

  

 

#2 Identify VMs with Old and Large Snapshots

A VM snapshot is a copy of the virtual machine disk file that preserves the disk file system and memory of the VM, and enables you to revert to the snapshot as a rollback or disaster recovery mechanism. While snapshots can be real lifesavers when upgrading or patching applications and servers, they also consume a lot of disk storage and even end up choking the performance of the VM. This is a typical problem caused by VM sprawl.

 

Look for VMs with old snapshots, large snapshots and also a large number of snapshots. Identify which are the snapshots are really required for business continuity and delete the rest to reclaim storage space for your VMs.

 

 

#3 Discover Orphaned VMs

An orphaned virtual machine is one that was created initially but, over time, has lost its association with its host. A VMs also shows as orphaned if it exists on a different host than the one expected by the hypervisor. This is a cause of VM sprawl as these VMs have been allocated resources in the past, but you do not have the visibility into which host is linked to them. Orphaned VMs are not really idle. The consume memory and CPU cycles, disk capacity, and add complexity to data protection.

 

 

#4 Monitor Oversized and Undersized VMs

An oversized VM is one that consistently uses less capacity than its configured capacity. Oversizing a VM you can result in decreased performance of other VMs inside the cluster. Conversely, an undersized VM is one that runs more load than what its configured resources can support. This can hurt VM performance and cause VM resource hogging that causes contention with other VMs in the cluster. You need to constantly look for the statistics of oversized and undersized VMs in your virtual environment and right-size the VMs by reclaiming and providing resources where they are needed most.

 

These are only some indicators to identify VM sprawl. To keep VM sprawl at bay you need to get visibility into the resource allocation and utilization of the hosts and VMs constantly and monitor VM performance metrics over time to understand the growth trend of VMs being added.

 

  

 

Some Tips to Keep VM Sprawl Under Control

  • Implement a formal process for requesting new virtual machines that requires justification for requests for any new VMs.
  • Monitor VM lifecycles so VMs can be deleted when no longer needed.
  • Actively monitor VM resource usage and get metrics on host resource usage so you can help the business understand the cost of virtual machines
  • Use a VM monitoring solution that provides visibility into VM usage, and allows you to right-size and reclaim over-provisioned resources, and delay new hardware purchases by optimizing the existing virtualized infrastructure for better performance.

 

For Additional Reading on VM Sprawl:

  • Watch this webcast to understand more about VM sprawl and how you can keep it under control