Zombie VMs are feasting on your data center’s valuable computing resources and costing you money. Zombie VMs, virtual machines that have not delivered data or compute power for six months or more, can negate the cost savings and operational efficiencies benefits of virtualization. Research shows that nearly 30% of virtual machines in data centers worldwide are zombie VMs.
They may be sitting idle, but zombie VMs still have storage, memory, and CPU allocated to them. They also contribute to licensing, power, and other operating costs while offering no value in return. Moreover, zombie VMs are often misconfigured and may not be updated with the latest security patches, which increases your network’s attack surface. It is time to zombie-proof your data center to enable efficient resource consumption and eliminate virtualization sprawl.
The ease and quickness of adding new VMs to a data center is a significant contributor to VM sprawl and the rise of zombie servers. At some organizations, new VMs are routinely added with no plan or schedule for decommissioning them. Often the person responsible for spinning up a VM moves on to a new role or leaves an organization before the VM has outlived its usefulness. Unless you have a robust asset management system in place, with detailed records about why each VM was provisioned and what it is being used for, it can be difficult to determine if or when they should be retired or moved from a development to a production environment.
It is tempting to let old VMs hang around your data center in case they are needed at some point in the future; especially if you don’t know what they were created for in the first place. However, if left unchecked, the spread of zombie VM’s can reach a point where they start degrading overall performance and make it difficult for administrators to effectively manage your data center infrastructure.
Hunting Zombies in Your Data Center
Finding all the zombie VMs in your data center can be a challenge. As infrastructures get bigger and become more complex, zombie VMs get harder to find. Key indicators to watch for include:
- Resource Usage: Monitor VM statistics, such as average CPU, RAM, and disk usage; network transmission rates; system uptime and heartbeat. VMs that are consuming no resources should be investigated further. It is important to note that some of your most critical VMs are used relatively infrequently. For example, a software license server may appear to be idle until new client machines are provisioned.
- Unpatched Services: Unpatched or misconfigured software and services may indicate the presence of a zombie. The lack of routine and periodic updates are a sign that the VM doesn’t have an owner and may not have a purpose.
- Server Logs: Check server logs for the last time anyone logged in or accessed the server remotely. VMs that haven’t been accessed in months might have been abandon and are good candidates for further investigation.
Once you have identified the VMs in your data center that are likely to be zombies, disconnect them from the network to see if users are affected before killing them off completely.
Combating Zombies with Hyper-Converged Infrastructure
Preventing VM sprawl and the rise of zombie servers in data centers is one of the primary reasons organizations are moving to hyper-converged infrastructure (HCI). Hyper-convergence combats unused and idle resources while reducing the cost and complexity of your environment by pooling virtual resources. With your entire data center storage and compute infrastructure combined in one solution you can get a holistic view of your entire environment using a single application. There are also full suites of software that you can use to monitor hyper-converged environments with tools that provide trending analysis and heat maps to help you get a handle on zombies before they become a problem. Hyper-convergence offers a host of benefits beyond killing zombies including backup and restore functionality, snapshot capabilities, data deduplication, inline compression, WAN optimization and more.
Put an End to the Zombie Apocalypse
Zombie prevention requires both policies and software tools. Your organization should have documented policies in place to guide the creation and retention of VMs in your data center, and software for monitoring their usage and reporting on how long their resources have been idle. Being vigilant, monitoring usage, and auditing your VMs on a routine basis will help you kill off the zombies in your data center.
Don’t wait for the zombie apocalypse before getting started; the longer you wait, the larger the problem becomes. If you require assistance identifying and eradicating zombie VMs, or if you are considering moving your data center to a hyper-converged architecture, Cerium can help. Our data center experts can put an end to money-sucking zombie VMs, recover their licenses, and harden your data center against cyber-attacks.