So today I took awhile to sort out some FT-related questions.
Some questions asked are… what does the 8 vCPU mean when someone purchases the vSphere Enterprise +? Does it mean the primary VM can be configured 8 vCPUs or does it mean both the Primary and the Secondary must be configured to a total of 8 vCPUs (meaning 2 x 4 vCPU).
I think first and foremost, to really understand how this all works, you’ll need to understand what the configuration maximums are.
For the sake of this blog post, I’ll screenshot the relevant part:
Fault Tolerance Maximums | Source
So first thing first, if you are on Enterprise Plus licensing, you are allowed to configure a maximum of 8 vCPUs per VM. This means, your primary VM can have a max of 8vCPUs configured for FT, and likewise, the secondary VM would have a max of 8 vCPUs.
However, since one host can only have a maximum of 8 vCPUs that are used for FT, it means that if you have fully configured an FT-VM (with a max of 8 vCPUs), you can only host one of such in that single host.
In certain situations where you configure a smaller FT-VM, say 4 vCPU, you are allowed to run 2 Primary VMs with 4 vCPU each configured for FT.
Take note that the passive/secondary VM can never be hosted on the same host as the primary – for obvious reasons. It basically defeats the purpose of having a fault tolerant VM.
So can we configure 8 FT-VMs x 1 vCPU and host them all on a single host? The answer is no- one host can only have a maximum of 4 VMs per host that are configured for FT. Hence, you can only host a maximum 4 FT-VMs x 1 vCPU each.
Hope this clarifies!