The vSphere Fault Tolerance feature does not support more than one vCPU. Why is support for multiple vCPUs in Fault Tolerance so desirable?
Fault Tolerance (FT) is a feature that delivers the highest level of availability in a VMware environment. FT gives uninterrupted service even if a VM fails.
But this protection comes at a cost. To supply constant uptime, FT requires at least two virtual machines (primary and secondary) running on two hardware ESXi servers. In addition to the added cost required to double the amount of host resources, a business needs to supply additional networking to enable FT. Due to the cost of additional hardware, most enterprises do not want to implement FT on all the virtual machines in the environment.
FT is ideal for VMs that are business critical, such as Exchange mail servers or database servers. Those workloads typically run with more than one vCPU to handle the heavy usage demands. Because FT still only supports a single vCPU virtual machine, the only alternative is to configure VMware High Availability, which results in a short downtime if a server fails and a VM is restarted.
Multiple vCPU FT exists, but it has been mired in a technical preview state for quite some time. There has been no announced date for when it will be available in vSphere, but there is speculation it will be released some time in 2014.
This was first published in December 2013