vSphere High Availability, Fault Tolerance and Snapshots

Question

I've heard claims from a vSphere server administrator that VMs configured for high availability continue to transparently function even in the event of an ESXi host failure. He also claims VMs configured for high availability cannot be snapshoted because that causes VM corruption.

To the best of my understanding, and from what I found online that's not true.

1. HA (high availability) VMs are restarted on failure, they don't transparently continue to function.
2. FT (fault tolerant) VMs transparently continue to function.
3. You can snapshot an HA VM but not an FT VM.

So, I'm a bit confused. What exactly are the benefits and limitations of HA and FT?

Answer 1

In summary, you're right and he's wrong; HA is much simpler, FT is very strict.

VMware High Availability (HA)

Restart virtual machines on other vSphere hosts in the cluster without manual intervention when a server outage is detected.

http://www.vmware.com/uk/products/vsphere/features/high-availability

An individual virtual machine can't be set for High Availability. High Availability (HA) is something you configure at the level of a cluster, not an individual virtual machine. Because of that, HA has no particular involvement with snapshots or disk provisioning models.

VMware Fault Tolerance (FT)

continuous availability for applications in the event of server failures by creating a live shadow instance of a virtual machine that is always up-to-date with the primary virtual machine. In the event of a hardware outage, vSphere FT automatically triggers failover—ensuring zero downtime

http://www.vmware.com/uk/products/vsphere/features/fault-tolerance

Fault Tolerance Snapshots

Note: As taking snapshots of FT virtual machines is not supported

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1016619

Fault Tolerance and Thin Provisioning

The shared virtual disks must be in the eagerzeroedthick disk format to facilitate clustering configurations, such as Microsoft Clustering service and VMware Fault Tolerance.

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1033570

But it seems like you can enable FT on a VM with thin provisioned disks, but it will cause the disks to be changed and that will take time and disk space.

Be aware that if you turn on FT on a VM, and that VM is using a thinly provisioned VMDK, that VMDK is inflated with zeroes.

http://blogs.vmware.com/vsphere/2012/03/thin-provisioning-whats-the-scoop.html