Microsoft Enhances Hyper-V R2 for Failover Clustering
Hyper-V R2 now supports a maximum of 1,000 virtual machines (VMs) in a cluster. Previously, in May, the hypervisor just supported a total of 960 VMs per cluster. A cluster can be defined as a grouping of host machines and their VMs that are managed as one entity. Typically, there would be at least two physical computers in a cluster, with one serving as the master (or job scheduler) and the other as the slave (performing the work).
The increased capacity was enabled because Hyper-V R2 now supports up to 384 VMs per node. Microsoft described the earlier limitations in a blog post on Wednesday.
"Until May 2010, the support limit was 64 VMs per host and up to 16 nodes in a cluster with one node for failover (a 15+1 cluster) for a total of 960 VMs," the blog explained.
Users of Hyper-V R2 -- either the standalone server or the one rolled into Windows Server 2008 R2 -- can vary the number of VMs per node for a maximum of 16 nodes per cluster. Microsoft supplies a table illustrating the possible number of VMs per node in its blog post.
Microsoft recommends reserving at least one node as a failover for the cluster. This failover node, or "passive node," stays idle, waiting to take over work should another node fail. With one node reserved for failover, a typical cluster may leverage a maximum of 15 nodes.
Microsoft further cautions that the number of VMs that can be run at the same time depends on three factors:
- "Amount of physical memory being used by each virtual machine.
- "Networking and storage bandwidth.
- "Number of disk spindles, which affects disk I/O performance."
Microsoft describes the requirements for running VMs using Hyper-V, as well as the limitations, in a TechNet article here.
Kurt Mackie is online news editor for the 1105 Enterprise Computing Group.