Error: "The host does not have sufficient CPU resources to satisfy the reservation. The host #### cannot satisfy the requested CPU resources of #### MHz." using vSphere High Availability after a physical host failure.
Article ID: 441475
Updated On:
Products
Issue/Introduction
- Unable to power on Virtual Machines (VMs) using HA after a physical host failure.
- Cluster uses Strict Admission Control with Reserved failover CPU/Memory capacity enabled.
- The impacted VMs typically have a 100% CPU reservation.
Environment
vCenter: 7.0U3C
ESXi: 7.0U3C
VCF: 4.4
TCP: 2.2
Cause
- This issue occurs because the physical loss of a host exceeds the configured vSphere HA admission control buffer.
- In a 16-host cluster, each host represents 6.25% of total resources. If your policy is set to a 6% reserve, the loss of one host (6.25%) consumes the entire safety budget.
- Under a Strict admission policy, vSphere HA blocks the power-on of high-reservation VMs to prevent a violation of the configured resource safety margin.
Resolution
To restore VM operations while the failed hardware is being repaired, you must perform one of the following:
Temporarily reduce the Reserved failover CPU/Memory capacity from 6% to a lower value (such as 0% or 3%). This releases the locked resource buffer and allows the cluster to utilize remaining capacity on surviving hosts.
Note: Lowering the Admission Control threshold to 0% means the cluster will not have a guaranteed resource buffer to restart VMs if a second host fails during the hardware repair window.Lower the CPU reservations on the affected VM from 100% to a lower threshold (e.g., 80%). This reduces the resource footprint required for the HA mathematical validation to pass.
Note: This may introduce "CPU Ready" latency if the cluster becomes overcommitted.Replace or repair the failed host to restore the necessary physical capacity to the cluster.
Additional Information
Feedback
