TL;DR
A Hyper-V cluster runs the Hyper-V role on every node of a Windows Server Failover Cluster, pooling hosts’ compute, memory, and storage into one unit.
- Requires the Hyper-V role, Failover Clustering feature, shared storage, and a common AD domain
- Cluster Shared Volumes (CSVs), typically NTFS, let every node read and write the same LUN
- Planned failovers live-migrate VMs before maintenance; unplanned failovers restart VMs automatically after a host fails
When organizations design their virtualized infrastructure for high availability, more than one Hyper-V host is needed to ensure that the workloads can withstand hardware failures, maintenance operations, and other events. While a single Hyper-V host allows running multiple virtual machines, running a Hyper-V cluster is generally required for production for availability, pooling resources, and taking advantage of the enterprise features offered by Hyper-V.
Let’s look at what is a Hyper-V Cluster for beginners and see how a Hyper-V cluster functions to serve virtual workloads.
What is a Hyper-V Cluster?
As we discussed in the previous article – Hyper-V Shared Storage for Beginners, having resources that only one Hyper-V host can access leads to increased risk, as any failure of the single host will take down all the resources associated with the single host.
A Hyper-V cluster is a Windows Server Failover Cluster (WSFC) type where the Hyper-V role is highly available and hosted on each node in the Windows Failover Cluster configuration. In this configuration, the Hyper-V servers in the Hyper-V cluster have shared storage and network connectivity needed to provide compute, storage, and networking for virtual machines.
The Hyper-V cluster functions as a single logical unit comprising multiple Hyper-V hosts that contributes to the computing, memory, and storage to the cluster. Each Hyper-V cluster host contributes to the overall resources available to workloads running in the cluster.
Hyper-V clusters are generally designed with headroom above the resources needed for the current number of virtual machines so the cluster can withstand the failure of one or more hosts. However, the Hyper-V cluster will still have the resources required to run all critical workloads on the remaining Hyper-V cluster nodes in this design.
Network requirements
There are networking best practices to consider while implementing a Hyper-V cluster. The basic network configuration in a simple Hyper-V cluster will generally have public and private networks. The public network configuration is configured on each Hyper-V cluster host allowing cluster connectivity from clients and connectivity to and from virtual machine traffic.
The private network is a separate network configured on the Hyper-V cluster hosts used by the cluster for the special heartbeat traffic required for Hyper-V clusters. It is essential to think about and design the networks configured in the Hyper-V environment to be unique, as overlapping subnets can cause issues with connectivity and routing if these are enabled later.
You also want to ensure that DNS is configured correctly and that each Hyper-V host in the cluster can resolve the other Hyper-V hosts for name resolution. You will also want to ensure all servers in the Hyper-V cluster are joined to the same Active Directory domain.
Roles and Features requirements
Creating a Hyper-V cluster requires installing the Hyper-V role on each Hyper-V host participating in the Hyper-V cluster. Each cluster host must have the Hyper-V role installed/ enabled to act as a hypervisor host. This requirement is no different than if it were a standalone Hyper-V host.
In addition, the hosts in the Hyper-V cluster also need to have the Failover Clustering feature installed on all hosts in the Hyper-V cluster.
Failover Clustering is the Windows feature allowing the creation of a Windows Server Failover Cluster that will host virtual machines in a highly available way.
Cluster Shared Volumes (CSVs)
Cluster Shared Volumes (CSVs) are the type of shared storage in Windows Server Failover Clusters (WSFC) and Azure Stack HCI that allow multiple hosts to read and write to the same LUN at the same time. The LUN can be formatted as NTFS or ReFS. However, there are advantages to using NTFS, as ReFS has some performance limitations with Cluster Shared Volumes.
CSVs allow the cluster nodes to failover quickly between each node in the cluster without requiring a change in drive ownership or other operations, such as dismounting/remounting a drive. The Cluster Shared Volume is the preferred type of shared storage when sharing storage between Hyper-V hosts as it provides the most benefits and capabilities for running workloads.
Failover – Planned and Unplanned
The key to a Hyper-V cluster is something called a failover. A host fails to continue contributing to the resources and capacity of the Hyper-V cluster. Failover can be planned and unplanned. The whole idea behind a failover cluster is to withstand failures. The act of failing over from one Hyper-V host to another is the failover process.
In a planned failover, such as when a host needs patches installed or during other maintenance operations, the workloads running on the Hyper-V host are intentionally drained from one host to another. It means the virtual machine workloads running on the host are Live Migrated from the Hyper-V host that will be taken down and moved over to the other hosts running in the cluster. Once the resources are migrated off the Hyper-V host, it can be taken down for maintenance operations, and business-critical workloads continue running as intended.
In an unplanned failover, if something bad happens, such as a hardware failure. It is unexpected, and so the failover is unplanned. If a Hyper-V host blue-screens and crashes, all the virtual machines running on the Hyper-V host will go down. However, the virtual machines will be restarted on a healthy host running in the Hyper-V cluster. While the failover was unplanned, the Hyper-V cluster performs as designed and can recover the workloads that went down during the outage automatically.
Hyper-V Cluster FAQs
What is a Hyper-V Cluster?– It is a special Windows Server Failover Cluster that hosts Hyper-V as a high availability role. A Hyper-V cluster combines multiple Hyper-V hosts and aggregates each host’s compute, memory, and network resources into a single logical resource for virtual workloads.
What are Cluster Shared Volumes (CSVs)? – Cluster Shared Volumes are special shared storage volumes that provide many advantages in terms of performance and efficiency in a Hyper-V cluster. CSVs can be formatted with NTFS or ReFS and allow multiple hosts with read and write access to the same storage simultaneously.
What is a failover?– A failover event happens when a cluster host fails to continue providing resources to the Hyper-V cluster. Failovers can be planned or unplanned. Planned failover happen when admins take down a host for maintenance, patching, or other upgrades. Unplanned failover generally happen due to hardware failover, power events, or other unplanned or unforeseen factors.
Wrapping Up
Hopefully, this general overview of a few of the concepts regarding Hyper-V clusters will help beginners to have a good idea of what Hyper-V clusters are and what purpose they serve. In production virtualization environments, clusters are generally required due to the availability requirements of most business-critical workloads. In addition, a Hyper-V cluster is fairly easy to configure. We will take a look at this process next.
Related Posts:
Hyper-V Guest Clustering – Part 2 – Steps to Create a Hyper-V Guest Cluster
Beginner’s Guide for Microsoft Hyper-V: How to Create Hyper-V Cluster – Part 15
Hyper-V High Availability and Failover Clustering
Hyper-V Cluster NIC Teaming
FAQ
1) What is a Hyper-V Cluster?
A Hyper-V cluster is a Windows Server Failover Cluster (WSFC) where the Hyper-V role runs as a highly available role on every node. It combines multiple Hyper-V hosts’ compute, memory, and storage into a single logical resource for virtual machines.
2) What software components does a Hyper-V cluster require?
Each host needs the Hyper-V role installed, just like a standalone host, plus the Failover Clustering feature. Failover Clustering is the Windows feature that lets a group of hosts form a Windows Server Failover Cluster capable of hosting VMs with high availability.
3) What are Cluster Shared Volumes (CSVs)?
CSVs are shared storage volumes in a Windows Server Failover Cluster that let multiple hosts read and write to the same LUN simultaneously. They can be formatted as NTFS or ReFS, though NTFS is generally preferred since ReFS has some performance limitations with CSVs.
4) Should Cluster Shared Volumes use NTFS or ReFS?
NTFS is the preferred format for CSVs, since ReFS carries some performance limitations in Cluster Shared Volume scenarios. CSVs let cluster nodes fail over quickly without dismounting or remounting drives or changing drive ownership.
5) What network setup does a Hyper-V cluster need?
A basic Hyper-V cluster uses both a public network, for client and VM traffic, and a private network, dedicated to cluster heartbeat traffic. Subnets should be unique to avoid connectivity and routing conflicts.
6) Do all hosts in a Hyper-V cluster need to be in the same domain?
Yes. Every Hyper-V host in the cluster should be joined to the same Active Directory domain, and DNS must be configured so each host can resolve the others for name resolution.
7) What is a failover in a Hyper-V cluster?
A failover happens when a cluster host stops contributing resources to the cluster, either in a planned event like maintenance or an unplanned event like a hardware failure. Failovers can be planned or unplanned, and the cluster is designed to withstand both.
8) What happens during a planned failover versus an unplanned one?
In a planned failover, VM workloads are live migrated off a host before it’s taken down for patching or maintenance. In an unplanned failover, such as a host crash, the affected VMs restart automatically on a healthy host in the cluster.
Follow our Twitter and Facebook feeds for new releases, updates, insightful posts and more.
