ClusterFailoverLevelAdmissionControlPolicy
The ClusterFailoverLevelAdmissionControlPolicy defines the number of host failures that should be tolerated and still guarantee enough unfragmented resources to failover all powered on virtual machines on those failed hosts.
When you use the failover level policy, vSphere HA partitions resources into slots. A slot represents the minimum CPU and memory resources that are required to support any powered-on virtual machine in the cluster.
With the failover level policy in place, HA uses the following slot calculations to control virtual machine migration within the cluster:
Calculate the slot size from CPU and memory reservations. The CPU value is the largest CPU reservation for all powered-on virtual machines in the cluster. The memory value is the largest memory reservation (plus memory overhead).
If your cluster contains any virtual machines that have much larger reservations than the others, they will distort slot size calculation. To avoid this, you can specify an upper bound for slot sizes; use the configuration editor in the vSphere Client to set the das.slotCpuInMHz and das.slotMemInMB attributes. When you use these attributes, there is a risk that resource fragmentation will cause virtual machines with resource requirements larger than the slot size to be assigned multiple slots. In a cluster that is close to capacity, there might be enough slots in aggregate for HA to successfully failover a virtual machine. However, if those slots are located on multiple hosts, a virtual machine assigned multiple slots cannot use them because a virtual machine can run on only a single host at a time.
Determine how many slots each host in the cluster can hold. HA uses the CPU and memory resources in a host's root resource pool to determine host slot capacity, not the total physical resources of the host. Resources used for virtualization purposes are not included. HA uses connected hosts that are not in maintenance mode and that do not have any HA errors.
The CPU slot resource is the host CPU resource amount divided by the CPU component of the slot size; the result is rounded down. HA makes the same calculation for host memory resource amount. HA compares the results; the lower of the two numbers is the host slot capacity.
Determine the current failover capacity of the cluster. This is the number of hosts (starting from the largest) that can fail and still leave enough slots to satisfy all of the powered-on virtual machines.
Compare the current failover capacity to the configured failoverLevel. If the current failover capacity is less than the configured failover level, HA disallows the operation.
{
"_typeName": "string",
"resourceReductionToToleratePercent": 0,
"pMemAdmissionControlEnabled": false,
"failoverLevel": 0,
"slotPolicy": {
"_typeName": "string"
}
}Number of host failures that should be tolerated, still guaranteeing sufficient resources to restart virtual machines on available hosts.
The base class ClusterSlotPolicy is used for specifying how the slot size is to be computed for the failover level HA admission control policy.
By default, vSphere HA defines the slot size using the largest memory and cpu reservations of any powered on virtual machine in the cluster. Subclasses of this class define various policies to modify how the slot size is chosen to prevent outlier virtual machines (i.e. those with much larger reservations than the average) from skewing the slot size. If such a policy is chosen, outlier virtual machines will use multiple slots. Using such a policy introduces a risk that vSphere HA will be unable to failover these virtual machines because of resource fragmentation.