Using VMware DRS and DPM to improve resource allocation

VMware DRS and DPM enlist vMotion to move VMs among physical servers for automatic resource allocation and load balancing. And VMware DPM can help save on data center energy costs.

VMware Distributed Resource Scheduler (DRS) and Distributed Power Management (DPM) improve resource allocation,...

efficiency and power consumption in virtual infrastructures.

VMware DRS balances workloads according to available resources, and users can configure DRS for manual or automatic control. If a workload’s needs decrease drastically, VMware DRS can temporarily power down unnecessary physical servers.

Working in a DRS-enabled cluster, VMware DPM reduces server energy consumption by consolidating virtual machines (VMs) across physical hosts. Both DRS and DPM use VMware vMotion to migrate VMs among physical servers. In conjunction with VMware High Availability, these features also help prevent server downtime.

The answers to these frequently asked questions will help you understand the basics of VMware DRS and VMware DPM and learn to configure them for more efficient resource allocation.

How exactly does VMware DPM work?                       

VCenter Server can suspend an idle server to save power. But only another host can restart a server, so VMware Distributed Power Management keeps at least one host powered on to send wake-up packets. You can configure VMware DPM on each host, only including those that you want to have standby capabilities. Once a server resumes, it boots as usual. Also note that Fibre Channel host bus adapters go down during standby, so storage admins might notice port drops on the storage network.

How does VMware DRS determine cluster balance and resource allocation?

To determine server load-balancing metrics, Distributed Resource Scheduler analyzes each host’s CPU and memory usage. When the standard deviation number of a host’s load is greater than that of the target host, DRS recognizes that a cluster is unbalanced . VMware DRS prioritizes which VMs should be migrated to restore balance, then uses vMotion to move VMs from the overloaded host to another. VMware DRS also re-evaluates the cluster every five minutes to check that the resource allocation does in fact restore balance.

How do I enable VMware DRS in a cluster?

When you build a cluster or anytime thereafter, you can set up VMware Distributed Resource Scheduler. Enabling VMware DRS occurs through a simple checkbox in the cluster’s settings. You also have to select a level of automation for VMware DRS. Many people choose Fully Automated, but you can choose Manual if you want to approve resource-allocation recommendations before they occur. You can also apply affinity rules, which place limits on where VMware DRS can move VMs.

What are the changes to DRS and DPM in vSphere 4.1?

The release of vSphere 4.1 improved VMware DRS clustering, especially its affinity rules. You can now group VMs and limit their migration to a certain set of hosts. This capability makes it easier to meet software vendors’ licensing terms, because many limit the migration of VMs from one host to another. As for VMware DPM, vSphere 4.1 added a scheduling task option that allows admins to apply DPM settings during or outside of production hours. In vSphere 5, due later this year, VMware will also add a Storage DRS feature for storage resource allocation.

What’s the difference between Hyper-V initial placement and VMware DRS?

Hyper-V’s initial placement feature selects the best host for a VM when the VM powers on. But unlike Distributed Resource Scheduler, the feature cannot move VMs after this placement. Hyper-V administrators must manually redistribute VMs when a host in standby comes back online, but VMware DRS does this task automatically. Initial placement in Hyper-V is a singular capability, but when used with Live Migration, it could perform some of the same actions as VMware DRS.

Dig Deeper on DRS and DPM