Zombies are everywhere, needlessly consuming resources in every customer’s environment. Zombie VM’s are virtual machines that were provisioned and for one reason or another unintentionally not used. This can be due to stalled or abandoned projects or even incomplete decommissioning procedures. Despite being unused, Zombie VM’s continue to eat valuable Compute, Memory, Storage, and even Network resources costing organizations money and exposing unnecessary security risks. What’s worse is that zombie VM’s have a negative impact on the environment as a result of the additional power and cooling required to feed these unused workloads.
VMware’s vRealize Suite is in a prime position to help customers identify Zombie VM’s. vRealize Operations can analyze metrics and properties of VM’s, applications, and physical hosts to identify wasted resources and associate a monthly cost with them. vRealize Log Insight can be used to validate that a VM or an application has not been accessed by analyzing system logs. And vRealize Network Insight can analyze traffic flows to and from your virtual machines and quickly identify VM’s with little to no network traffic. With these tools, organizations can identify and eliminate a lot of waste with very little effort.
This seemed like a great way to challenge our Cloud Management vExperts. For the uninitiated, the vExpert Cloud Management program is a sub-program of the greater VMware vExpert community. vExperts are hand-picked every year and awarded the highly sought-after title in recognition of their efforts outside of their day-jobs. vExperts are primarily customers who have a passion for VMware’s products and are willing to share their experiences through blog posts, presentations, and other publicly available spaces such as YouTube. The VMware vExpert program is about to launch into its 11th year and the vExpert Cloud Management sub-program will be entering its second year in 2020.
On November 15th, 2019 we kicked off our vExpert Cloud Management Zombie VM Hunt with VMware’s Director of Sustainability Innovation, Nicola Peill-Moelter presenting on the Zombie VM problem. She shared the economic and environmental impacts that zombies pose with our group. We then challenged our vExperts to share how they would use vRealize Operations, Log Insight, or Network Insight to identify Zombie VM’s. In return for their efforts, we offered each participant a custom vExpert Cloud Management golf shirt and a donation to plant trees in the Amazon Rainforest in their name.
Since the Zombie VM Hunt began, news of the Australian wildfires began to pop up as a major global issue. According to CNN, these fires have destroyed thousands of homes, killed nearly 30 people, and impacted an estimated half a billion animals. As of this blog’s writing, the fires continue to burn due to extreme droughts and have burned more than 17.9 million acres. Because of the incredible scope of this devastation, it only seemed right to instead donate to the replanting of trees in Australia through the One Tree Planted organization. I’m happy to share that 60 new trees will be planted in Australia as a result of our vExpert’s efforts.
Our Cloud Management vExperts took to the challenge and shared their ideas and experiences via blog posts, custom dashboards, and even presenting the challenge and solutions at a local VMware User Group (VMUG) meeting in Myanmar!
In this first blog post, Graham Barker explains how he used vRealize Operations to identify zombie VM’s. By analyzing the age of Virtual Machines, the uptime, virtual hardware and VMware Tools versions, as well as network utilization, Graham was able to identify Zombie VM’s and bundled it all into a single dashboard that you can use today!
https://virtualg.uk/using-vrealize-operations-to-identify-zombie-vms/
Matt Menkowski shows us how to create a custom view to identify powered off VM’s. While powered off VM’s don’t consume physical CPU or Memory, they still consume storage. Be sure to click on the two tabs at the top of Matt’s blog post. The first shows you what is available out of the box with vRealize Operations, and the second tab walks you through the custom view creation. There’s also a URL to download the view and use it right away.
https://vmscribble.com/vrops/vrops-8-vm-powered-off-timestamp-view/
Alain Greenrits shared a custom vRealize Operations view to identify idle VM’s. By identifying and color coding VM’s based on CPU, Memory, Disk, and Network usage. VM’s that are highlighted green indicate they are using less than 500 MHz of CPU, 500 MB of memory, and 100 kBps of network and disk throughput. He also incorporated this view into a custom dashboard that you can download via the link in his blog.
https://bitstream.geenrits.net/vmware/detect-historical-idle-vms-on-vrops/
And finally, Elizabeth Souza created a custom vRealize Operations dashboard to show idle and powered off virtual machines, datastores with wasted space, and even oversized VM’s.
Elizabeth’s blog is available here…
http://geeklady.com.br/index.php/2020/01/04/procurando-por-vms-zumbis/
And you can download her dashboard here…
https://code.vmware.com/samples/6753/identifying-zombies-vms?h=vRealize%20Ops%20Dashboard
We want to thank each of our Cloud Management vExperts for taking the time to share their solutions to the Zombie VM problem with the rest of the community. We hope you will find their ideas useful in using the vRealize Suite to save your organization from waste and reduce carbon emissions generated by keeping unused VM’s running. For more information on VMware’s vExpert program, visit https://vexpert.vmware.com/. For more information on VMware’s vRealize Suite, visit https://www.vmware.com/products/vrealize-suite.html.
Special thanks to Nicola Piell-Moelter and Steve Tilkens for making this program possible!!
