Servers are essential when running a company, but it’s easy for those resources to sprawl. Server sprawl comes with having too many servers, which results in unnecessary complexity and inefficiency when those servers aren’t used to capacity.
The issue doesn’t go away with virtualization, either. Virtual machine sprawl, or VM sprawl, is a problem for many businesses. In some ways, the ease of using virtual machines increases the likelihood of sprawl.
This problem has a solution, though. By taking certain steps and using a tool like SolarWinds® Virtualization Manager (VMAN), you can streamline processing, reduce sprawl, and become more efficient.
What Is VM Sprawl?
Impacts of VM Sprawl
Managing VM Sprawl
Use a VM Management Tool
Containing VM Sprawl
What Is VM Sprawl?
VM sprawl, also known as virtualization sprawl, happens when an administrator can no longer effectively control and manage all the virtual machines on a network.
This can happen with rapidly growing networks when multiple VMs are set up for use by different departments. This use is often temporary—for example, an organization might set up VMs as part of a testing environment—but if the VMs aren’t archived when they’re no longer needed, administrators can end up with too much to manage.
Additionally, having these idle VMs means the network loses efficiency. Even when VMs are actively being used, they may not be configured for optimization, which reduces the benefits of virtualization.
Impacts of VM Sprawl
It’s important to understand why VM sprawl matters. VM sprawl impacts businesses in a few different areas.
The first issue is cost. Storage is less expensive than it used to be, but it still has a price, and when VMs are on, they use host storage, memory, and CPU time. On top of this, even virtual machines need licenses for the operating system and applications they’re running. This adds up to a substantial cost for something you may not even be using.
VM sprawl results in substantial management overhead. VM administrators need to manage both the physical resources and the VMs. Storage administrators have even more responsibility: they need to keep the VMs backed up regularly, create more volumes or datastores, and rebalance the storage. This takes up time those administrators could be using for other tasks.
Risk comes into the picture when administrators don’t have a good way to track the VMs. This makes it hard to identify which ones are being used for production and which were intended to be temporary. Another possible risk involves the person who knows the VM setup leaving the company without having created clear records of the configuration.
Lack of Space
Virtual machines don’t take up physical space, but they do consume a lot of disk space. When a single VM can take up multiple gigabytes, it doesn’t take long to use all available resources. This can lead to additional disk purchases that aren’t truly necessary and wouldn’t be made if the utilization were fully understood.
Most companies want to know what their resource usage is like, so they can prepare for necessary purchases or upgrades. With VM sprawl, you won’t have a complete understanding of usage, which prevents IT teams from accurately forecasting future usage and needs.
Performance is another area of concern, since having many VMs running puts a strain on storage and slows everything down. The impact is even more significant when the VMs run automated or scheduled tasks, including defragmentation and virus scanning. It’s possible to move VMs around to manage the capacity, but this still wastes resources.
Finally, VMs need to be backed up. This can be a major undertaking for backup administrators and having more VMs puts additional strain on the backup systems. Even worse, the effort is unnecessary if it’s spent on backing up idle VMs. Similarly, in the event of disaster recovery, more time and resources are wasted by restoring unused VMs.
Managing VM Sprawl
To avoid the many negative impacts of VM sprawl, it’s important to implement management practices that give you VM sprawl control. You won’t be able to eliminate sprawl completely, as you’ll still need to create VMs for different reasons, but with VM management, you can use VMs without being negatively impacted by sprawl.
Perform an Audit
The first step is to perform an audit of the VMs. This allows you to identify which are associated with a hypervisor cluster or host.
You can do this by running scripts to compare the inventory to the on-disk contents—if you find any machines not mapped to a hypervisor, you’ll know those aren’t in active use. You may then be able to delete those VMs to free up space.
Rescue Idle Resources
It’s also important to check for idle VMs with high co-stop percentages due to scheduling contention. These VMs are running, but since they’re not performing any tasks, they consume resources when they don’t need to. Once identified, you can decide whether they should be repurposed for another use, be shut down, or have their provisioning reduced.
Find the Right Size for VMs
A key element to controlling sprawl is to right-size your VMs. If VMs don’t have enough resources or have more than they need, adjusting their configurations will improve performance. This can be done manually or through a virtualization monitoring tool.
Clean Up Orphaned Snapshots
As part of VM management, it’s important to take snapshots to create a copy of the VM disk file at a specific point in time. These snapshots provide a log of changes and can be used to restore the VM.
You should have multiple snapshots, but over time these can become excessive. You don’t need to keep every snapshot, especially orphaned ones. Orphaned files remain even after the original VM is deleted. Orphaned and redundant files take up significant resources and cleaning them up frees disk space for other uses.
Use a Good Naming Convention
You might already have naming standards in place to help keep track of servers, but perhaps those conventions could be improved. For example, including the name of the business owner or department in the name or description of the VM, in addition to specific contact information, makes it easier to find the owner of a VM. If you notice problems with performance or see the VM is idle, you know who to contact.
Deleting VMs frees up resources, but it’s not always the best approach. In some cases, a VM might not be used now but will be needed again in a month or two.
In those situations, the best approach is to use archive functionality, which is often provided with backup products. This allows you to archive the VM elsewhere until it’s needed. This includes keeping track of where the VM is stored, so you can find it later, and cleaning up the archive if those files are no longer necessary.
End Unnecessary Backups
Backups are important, but backup repositories can also run out of space. When this happens, you should check the backups to see if any VMs are part of multiple backup jobs. Identifying these situations and removing the redundancies is another way to reduce sprawl.
Clean Up Junk
Everyday use of VMs is essential for many companies, but this constant use also generates a lot of garbage files. For example, you might still be storing temporary files and VM configuration data for VMs that have been deleted or are no longer in use. Identifying these junk files and removing them frees up resources for better uses.
One of the biggest challenges with VM management is keeping track of all the objects in the virtual infrastructure. If each VM is assigned a certain category, it becomes much easier to group and manage them.
You might have some VMs that don’t fit into neat categories, but many do. You can assign these categories based on the type of usage, department, project, etc. You can also define expected maximums for each category, which can help identify any outliers.
This type of approach also makes it easier to review the different VMs, since you can get all the information about a specific category in a single glance.
Establish Processes and Standards
You can also manage VM sprawl using certain processes and standards for the IT team.
The first step is implementing role-based access control to define the roles and permissions for different users. These policies ensure the right people can create VMs and snapshots, which will reduce the chances of having rogue VMs.
These processes could also include:
- Asking for justification on any oversized VM requests
- Adjusting resource allocation based on real utilization
- Adding guidelines for how many snapshots to keep for a given VM
Having these standards in place will make it easier to manage VMs and reduce sprawl.
Use a VM Management Tool
Finally, you can use a VM management tool to address virtualization sprawl. These tools help you make the most of a stack virtual machine and your whole VM network, by automating certain tasks and giving administrators clear visibility into the operation of VMs.
I recommend SolarWinds Virtualization Manager (VMAN). This solution has all the features needed to manage virtual machines, including VM sprawl control, recommendations for VM management, performance monitoring, customizable alerts, and in-depth dashboards. It meets all the requirements for an effective VM tool, as outlined below.
A good VM tool will have the following capabilities:
One of the first things to look for in a management tool is a way to monitor performance, because manually identifying slow performance for multiple VMs is inefficient, impractical, and subject to error.
With a VM performance monitoring tool, the process can be automated. Some of the metrics the tool should track are:
- VM snapshots, especially outdated ones
- Distribution and usage of VM resources
- Availability of datastore resource and capacity
- Memory issues
- Hardware health
The tool should also present this information in a dashboard to provide a central location for analysis and troubleshooting.
It helps to have a way to define alerts, so the administration team doesn’t have to constantly monitor the results. Instead, they can configure alerts for bottlenecks related to memory, CPU, and storage performance, with customizable thresholds to meet the needs of your business. These alerts keep administrators up-to-date with real-time information.
Virtual Object Management
Performance isn’t the only important information to look at. It’s helpful to have a tool to identify dependencies between virtual objects. Having visibility into those links between hosts, storage, clusters, and virtual machines makes it easier to manage the resources.
VM Sprawl Control
A good solution will also focus on issues of VM sprawl. With the right VM management tool, you can review the information in a dashboard to identify problems and optimize VMs. The dashboard might even offer recommendations about how to adjust CPU and memory to better allocate your resources, and it can help you reclaim idle VMs and manage snapshots.
VM Configuration Management
You may want to track VM configuration to identify how VM resources are deployed and how the settings are configured. VMs are quick to set up, but the default options aren’t always the best for optimization. Administrators often need to change those defaults to make the VM efficient.
Configurations aren’t constant, though. A VM might start out needing a certain amount of resources for best performance, but it can change over time. Additionally, sometimes configuration adjustments introduce problems instead of resolving them. Trying to manually review this information is time-consuming. With a good management tool, administrators can consult a history of configuration changes to easily find root causes and address the underlying issues.
Containing VM Sprawl
VM sprawl is a real problem, but by implementing best practices and using a VM management tool, you can make the most efficient use of your resources. With a tool like SolarWinds VMAN, VM sprawl is within your control. When administrators have easy visibility into VM performance and issues, along with ways to automate necessary changes, the whole network will run smoothly, and your business will see the benefits.
A 30-day free trial of VMAN can be downloaded here.