vSphere HA Component Protection events do not trigger despite APD occurrences

Products

• VMware vSphere ESXi
• VMware vCenter Server

Issue/Introduction

• vSphere HA is enabled, and VM Component Protection (VMCP) is configured.
• A path failure occurs on a target datastore, resulting in "flapping" (intermittent occurrences and recoveries of All Paths Down (APD) states).
• Despite these events, virtual machines (VMs) are not restarted.
• The following message is recorded in the fdm.log of the affected ESXi host:

  Datastore /vmfs/volumes/########-######## access failure cleared, and it is accessible now. Due to this accessibility change, # VMs recovered from datastore access failure

Environment

VMware ESXi 8.0
VMware vCenter Server 8.0

Cause

This is an expected behavior by design.
If an APD condition is cleared within the time period configured in the "Delay for failure response" setting after the initial occurrence, vSphere HA Component Protection will not trigger a failure response (such as restarting the VMs).

Resolution

Since this is the intended behavior of the product, there is no specific fix or workaround required for vSphere HA.
Please perform troubleshooting on the storage or network infrastructure to address the root cause of the intermittent APD (path failures).

Additional Information

Permanent Device Loss (PDL) and All-Paths-Down (APD) on host

The ESXi host experiencing latency and intermittent APDs on NFS datastores.

All Paths Down for a storage device

Japanese KB: APD が発生したにもかかわらず、vSphere HA のコンポーネント保護イベントが発生しない