VMware Fault Tolerance

What is Fault Tolerance ?

• FT is a technology used by VMware to provide zero downtime for a virtual machine
• Continuous protection for a VM even in case of a host failure
• FT is enabled only on a VM basis
• When enabled, FT creates a secondary VM in another ESXi
• Uses Record and Replay technology & vLockstep technology

Record and Replay technology

• This technology records all events in the primary VM and replays in the secondary VM( Even a mouse click) but only the primary VM produces output such as disk writes and network transmits
• It is done using a FT logging NIC
• The output of the secondary VM is suppressed by the ESXi and is not in the network until the primary fails
• If the primary fails, the secondary becomes the new primary

vLockstep technology

• This technology uses a second identical processor to monitor and verify the operations of the first processor
• Both processors receive the same inputs so the operation state of both processors will be identical
• Fault Tolerance feature relies on vLockstep technology to sync the CPU operations of a VM between two hosts so they are in identical states

Limitations

• FT requires newer processor models from AMD and Intel which supports vLockstep technology
• All hosts must be running the same build of VMware ESX
• Shared storage is required as a single vmdk is being used by two VMs in two separate hosts
• HA must be enabled
• VMs must be single-processor (till ESXi 5.x...From 6.0 this has been changed to 4 processors)
• All VM disks must be “thick”
• No devices like USB, physical CD-ROM, floppy, physical RDMs are permitted
• Snapshots must be removed before enabling FT
• Taking snapshot of a VM in FT is not possible
• Storage vMotion operation is not permitted for a VM in FT
• FT enabled virtual machine is automatically configured as DRS-disabled but initially the secondary VM is placed by DRS when FT is enabled