Notes for FT

In FT when both the VMs can be on same host?

Both primary and secondary VMs must be on separate host to avoid any single point of failure. But these VMs can come on same host when they are both in a power off state

What is primary requirement for FT enabled VM to work on DRS enabled cluster

EVC mode must be enabled for FT enabled VM’s to take advantage of DRS’s initial placement and migration recommendation

If EVC is not enabled how does it affects/limits FT enabled VM with respect to DRS

When vSphere Fault Tolerance is used for virtual machines in a cluster that has EVC disabled, the fault tolerant virtual machines are given DRS automation levels of “disabled”. In such a cluster, each Primary VM is powered on only on its registered host (i.e. where it was originally created), its Secondary VM is automatically placed (i.e. initial placement happens every time), but neither fault tolerant virtual machine is moved for load balancing purposes.

What is the maximum number of host DRS places on any single host?

DRS do not place more than a fixed number of Primary or Secondary VMs on a host during initial placement or load balancing. This limit is controlled by the advanced option das.maxftvmsperhost. The default value for this option is 4. However if you set this option to 0, DRS ignores this restriction.

How does VM-VM affinity and VM-Host affinity rules impact FT enabled VMs?

VM-VM affinity rule applies only to the primary VM. If VM-VM affinity rule is set on primary VM, DRS attempt to correct any violations that occur after a failover. While VM-Host affinity rule applies to both Primary and secondary VM.

Is vCenter required for FT to work?

No. FT enable machine will failover when the host fails. The failover of fault tolerant virtual machines is independent of vCenter Server, but you must use vCenter Server to set up your Fault Tolerance clusters.

List down the tasks which must be completed before enabling FT?

· Enable host certificate checking. This is by default enabled. Check this only if you’ve upgraded from vsphere4.0 or below.

· Create VMKernel port for FT

· Create vSphere HA cluster, add hosts and check compliance

What happens when you disable FT logging port when both primary and secondary VMs are running?

If you configure networking to support FT but subsequently disable the Fault Tolerance logging port, pairs of fault tolerant virtual machines that are already powered on remain powered on. However, if a failover situation occurs, when the Primary VM is replaced by its Secondary VM a new Secondary VM is not started, causing the new Primary VM to run in a Not Protected state.

What all checks are done before FT is enabled on a VM?

  • vSphere HA is enabled
  • SSL Certificate checking is enabled
  • Version of ESXi host is better or equal to ESXi4.x
  • VM has only one vCPU
  • VM has no snapshots
  • VM is not a template
  • vSphere HA is not disabled
  • VM do not have a device which is 3-D enabled

What all checks are done when FT enabled VM is powered ON

  • Hardware virtualization is enabled at the BIOS of ESXi host
  • Host’s processor must support FT
  • Host on which Secondary VM is to be created is FT compatible host and is of same processor family as host on which primary is being created.
  • Guest OS compatibility with FT
  • Hardware compatibility with FT
  • And checks for any unsupported device