Error: "vSAN is unable to retrieve the physical disk information from host" alert triggered in vSAN Health checks

Products

• VMware vSphere ESXi
• VMware vSAN 7.x

Issue/Introduction

• vSAN health alarms such as "Inconsistent configuration", "vSAN is unable to retrieve the physical disk information from host", or "CLOMD process issues" appear intermittently and may auto-resolve.
  
  
• The vSAN alert triggered are isolated to specific hosts and not on all hosts in the cluster.
  
  
•  vsanmgmt agent is unresponsive or fails to query storage information.
  
  
• In hostd.log, you see admission control failures:
  
    vim.fault.InsufficientMemoryResourcesFault

  Admission check failed for memory resource
  
  
• In vmkernel.log, you see warnings regarding ramdisk extension failures:
  
   WARNING: VisorFSRam: 213: Cannot extend visorfs file /vsantraces/###### because memory could not be allocated for its ramdisk (vsantraces).
  
  
• The ESXi host has sufficient physical memory capacity available and is not fully utilised. 
  
  

 

Environment

ESXi 7.0 U3. 
vSAN 7.0 U3. 

Cause

• The issue is caused by severe physical memory starvation on the ESXi host, typically triggered by aggressive memory reservations configured on local resource pools or specific virtual machines.
  
  
• When virtual machines or resource pools are configured with excessive reservations, it can starve the host below the recommended 6% overhead required for stable operations on high-memory hosts.
  
  
• The ESXi VMkernel strictly prioritizes to keeping the existing virtual machine workloads running by using the memory overhead required for the management daemons.
  
  
• Hence, though the ESXi host appears to run fine, the management daemons (like vsanmgmt and clomd) and system filesystems (visorfs)  are starved of memory as the host had no available unreserved memory over head.
  
  
• This leads vSAN daemons running out of memory, health query failures and the subsequent alarms.
  
  
  

 

Resolution

Ensure the ESXi host has sufficient unreserved physical memory 

1. Identify the host reporting the disk retrieval or configuration alarms from the vSAN health check.
   
   
2. Log in to the vSphere Client and navigate to the Hosts and Clusters view.
   
   
3. Locate the affected host or resource pool.
   
   
4. Right-click the resource pool, select edit resource settings, and reduce the reservation to a sustainable limit ( at least 6% of its physical RAM as unreserved overhead).
   
   
   • 512 GB - 1 TB hosts: Maintain a minimum of 32 GB to 64 GB unreserved memory.
     
     
   • 2 TB+ hosts: Maintain a minimum of 128 GB unreserved memory.
     
     
5. If high reservations are required for specific workloads, ensure vSphere DRS is enabled and set to fully automated. This allows the cluster to distribute reserved workloads across multiple hosts, preventing a single host from reaching a starvation state.