Has VM Sprawl left you feeling bloated?

Happy New Year!

Announcement: Have you seen the latest release of Foglight for Virtualization, Enterprise? It comes with many new features and improvements. Check it out at https://www.quest.com/products/foglight-for-virtualization-enterprise-edition/!


Virtual Machines are easy to provision and create. So IT processes to request a new VM are rather lenient. This results in VM sprawl. Users request new VMs and then use it temporarily. Usually the provisioned VMs show high use in the beginning, but many times, as the need passes, the use tapers. VMs are easy to create, but not so easy to delete for the IT staff.


In theory, this should not be much of a problem, since the VMs are sitting on shared resources and if the users stop using them, VMware platform can take care of taking resources away from them. Thus new VMs can be created on the same hardware.


But the VM sprawl catches up with IT at some point. More VMs mean more management and more resource contention. IT needs an easy to use tool to reclaim resources at are unused.

Foglight for Virtualization Optimization tool allows the IT staff to reclaim resources.

The Optimization tool allows IT to reclaim CPU and Memory resources by looking at actual use of the resources and comparing it to the allocated resources. So if a VM is using only 30% of CPU of the allocated 4 vCPUs, we can safely take 2 vCPUs out of that VM without limiting the performance of the guest.


Notice that the collected information needed to make the decision is all based on what Foglight already collects and does not need any input from the end user.


The right-sizing of the VMs serves two purposes: It allows IT to reclaim unused, but allocated resources and it simplifies provisioning of new VMs.


IT staff can create new VMs from a simple template and offer a set number of CPUs and memory to the VM. After observing the utilization (either the max utilization or average utilization) of the VM, IT can run optimization and reclaim only the unused resources, thus simplifying the provisioning process and still not having wasted resources.

Storage waste also occurs when IT staff copies VM data files in order to create new VMs etc. but only to abandon them. Another source of abandoned VMs is when VMs are removed from vCenter inventory, but not deleted from the VMFS. Optimization offers a tool to recover wasted resources from the above situations as well as find VMs that have been abandoned and are still consuming resources.

Optimization cartridge can create reports to help IT find abandoned VM images and zombies as well as powered-off VMs. Zombie VMs are powered-on VMs but no one is logging into them or they are usually not running any apps. So these VMs consume very small amount of compute resources as well as storage.


Using Optimization, IT can recover the storage and compute resources from the zombie VMs.

Another aspect of VM sprawl is accumulating waste over time. Users take snapshots and fail to remove them. VM snapshots accumulate space and consume new space on VMFS. Thus VMDK files continue to grow in size once a snapshot is taken until it is removed.


Using Foglight optimization, IT staff can find out the snapshot size and the time when it was created. Looking at it, the recommendation is generated to delete the snapshot.


In summary, VM sprawl costs real money and a tool like Optimization is needed to check it.