Virtual Machine (VM) sprawl (also known as VM server sprawl and VM image sprawl) is defined as the inability to control the growth of virtual machines in your environment. It is created when you provide VM provisioning access to users and allow them to create their own VMs.
With this in mind, a spike in the number of VMs created makes it difficult to determine which VMs should be maintained and deprovisioned. For this reason, administrators should implement a VM sprawl management process that prevents, controls and remediates that growth.
1. What is VM Sprawl Management?
VM sprawl management is a process that enables you to control VM sprawl throughout your virtual environment. First, managing VM sprawl is vital to protecting resources like memory and disk space. When additional applications consume memory and disk space, you inherently hurt your bottom line.
Furthermore, VM sprawl control is crucial for forecasting. For example, you may need to know exactly how many resources are required to service your infrastructure in six months. Controlling VM sprawl allows you to forecast and plan for hardware upgrades more effectively.
Generally, it’s not about stopping VM sprawl; your organization needs to create and delete virtual machines on a regular basis. Instead, it’s more about controlling VM sprawl and managing the VM provisioning process.
2. How Do You Remediate VM Sprawl?
To reduce VM sprawl, you must implement role-based access control (RBAC) and VM lifecycle reporting. First, get administrative control over the user who can create virtual machines. Once you have control, you can restrict who can build VMs and what types of VMs different authorized users can deploy.
3. How Do You Prevent VM Sprawl?
VM sprawl can be prevented by limiting the locations in which users can deploy VMs. Administrators can also run VM sprawl reports on a daily, monthly and yearly basis. VM sprawl reporting enables you to detect VM sprawl and inform you of the VMs created. It also lets you determine how much disk space is consumed.
Consider the example of a development group who is responsible for creating applications for Windows 10 desktops. As an administrator, you can build a template for Windows 10 virtual machines and assign your team access to locations for creating new VMs inside that template.
Other examples include local file storage locations like a local disk on a Hyper-V host, a network file share like an SMB share or cluster shared volume. VM provisioning templates limit VM sprawl by standardizing what types of VMs users can create and where they can deploy newly created VMs. To summarize, the development group is not just randomly creating VMs in your environment; you control who creates VMs, what VMs are created and where they are located.
4. How Do You Detect VM Sprawl?
Unlike a performance bottleneck, VM sprawl is not something you can detect, but it is something you can control. Your business requires VMs to be created and destroyed frequently. VM sprawl management controls the creation and destruction of VMs. It also gives you a clear understanding of the number of VMs you create and destroy on a regular basis. In other words, you can control VM sprawl, but you cannot stop it.
5. What VM Sprawl Tools Can You Use?
Furthermore, there aren’t many native VM sprawl tools that are currently available on the market. System Center Operations Manager (SCOM) requires you to run custom reports for the number of VMs created and destroyed each day, and there is no Windows-based out-of-the-box solution. System Center includes basic RBAC and self-provisioning (the ability to allow users to create their VMs). However, this is the fundamental problem of what causes VM sprawl. When you give people access to build new machines, and in some cases don’t delete them, you have hundreds of users continuing that process, which ultimately results in VM sprawl. By default, no native tool, including SCVMM, has out-of-the-box reporting.
6. What Steps Should You Take to Control VM Sprawl?
- Keep track of the number of the VMs created and deleted
- Identify VMs that are consuming resources but aren’t being accessed
- Look for VMs that aren’t consuming many resources like storage, network and CPU
- Identify users who created the VM in question and delete those that aren’t necessary
7. How Does 5nine Cloud Manager Prevent, Manage and Reduce VM Sprawl?
5nine Cloud Manager provides the reporting, lifecycle management features and role-based access controls needed to reduce VM sprawl throughout your entire virtual environment. The solution starts by offering details on the number of VMs created compared to the historic data. Not only does it include a full audit trail of every VM created but it also keeps track of who created the VM and when.
Additionally, 5nine Cloud Manager searches for zombie VMs, which are virtual machines not in use. Additional reports provide insights on IOPS and identify virtual machines that are misconfigured or have poor memory allocation (along with CPU percentages that may need to be adjusted).
5nine Cloud Manager identifies VM sprawl and Zombie VMs by determining the number of CPU cycles, disks and network bytes. Based on these three numbers, the solution determines if the VM is operational.
Cloud Manager prevents VM sprawl via RBAC and its data store. You can use your existing Active Directory groups or create custom groups for users. Begin by assigning your desired group a data store (which is just a location for data storage) along with location permissions for new VMs. You can also assign users to template storage locations where you can store templates and allow your admins access to create and save their templates. Next, 5nine Cloud Manager audits the entire VM lifecycle, including when a user creates or deletes a VM. The solution makes these details available within the software but they are also available via Syslog, for any external analytical systems like Splunk or Microsoft OMS.
5nine Cloud Manager writes a record to its jobs view every time a user completes an action. It makes the data available via export or in real time via Syslog. In addition to VM sprawl and Zombie VM reporting, 5nine Cloud manager also provides remediation actions. For the VM in question, users are given the option to delete or ignore it depending on its purpose. For incorrectly sized VMs or those virtual machines that have too little or too many CPU or memory cycles, users can “right size” appropriate resource allocations.
I am an author, speaker and technical evangelist focussed on Microsoft Cloud management and security. I’ve held product management and product marketing roles at early stage startups and enterprise software vendors, all with an emphasis on Microsoft technologies. As the Senior Evangelist for 5nine, I get to share the 5nine story with audiences all over the world. I talk, I blog, I record videos, and I spread the word via social media.