Home / Definitions / Failover Cluster

Failover Cluster

Forrest Stroud
Last Updated June 23, 2021 7:07 am

A failover cluster is a group of computer servers physical or virtual (or both) that are configured to provide continual access with no downtime or an absolute minimum of downtime. In a failover cluster, if one server, or node, stops working or needs to be rebooted, one or more other nodes will take over the existing workloads and also handle any new requests that come in.

As such, failover clusters are designed to provide either high availability (HA) or even continuous availability (CA) at all times, and are often deployed in enterprises to ensure access to mission-critical data and tech resources like databases, key applications, and shared resources.

High Availability vs. Continuous Availability Failover Clusters

High-Availability clusters (HA clusters) are designed to provide 99.999% availability (known as the five nines ), which equates to less than 6 minutes of downtime a year. CA clusters, or fault tolerant clusters, on the other hand, tout 100 percent availability and have more stringent requirements for preventing single points of failure, which is reflected in their heftier price tags in almost all cases.

In a Continuous Availability fault tolerant failover cluster, potential failure points are continuously monitored, from storage to processor units, power supplies, network components, to other hardware and software resources. As soon as any resource fails or shows signs of starting to fail, a fault tolerant failover cluster will switch over to a backup resource without any interruption in service or downtime.

Failover Cluster

Image Source: ServerWatch

Both High-Availability clusters and CA clusters are often deployed in conjunction with other types of clustering to provide load balancing, parallel processing, RAID storage, and similar capabilities.

Popular failover clustering solutions on the market today include Windows Server Failover Cluster (WSFC), VMware Failover Clusters (VMware HA and vSphere 6 Fault Tolerance), SQL Server Failover Clusters, IBM PowerHA SystemMirror, Linux-HA, Red Hat Linux Clusters, and Veritas Cluster Server.