Virtual machines become inaccessible after vSAN cluster restart.

Products

• VMware vSAN

Issue/Introduction

Symptoms

• Virtual machines (VMs) show as inaccessible in the vSphere Client or invalid in the ESXi Host UI following a planned cluster shutdown and restart.

• • vCenter Web Client:
    
    
    
    
  • Host UI:
    
    

• The vSAN datastore do not list any VMs.
  
  
  
  
• Running the recovery script fails with one of the following error messages:
  
  
  • python /usr/lib/vmware/vsan/bin/reboot_helper.py recoverChecking network status...Recovery is not ready, retry after 10s...Timeout, please try again later
    
    or 
    
    
  • python /usr/lib/vmware/vsan/bin/reboot_helper.py recover
Begin to recover the cluster...
Time among hosts are not synchronized. (drift>5s)
Time on localhost: 2026-09-22T10:27:21.619329+00:00
Time on ###.###.##.##: 2026-09-22T10:27:10.694486+00:00
Please enable NTP service across the hosts and vCenterServer, or synchronize the time settings manually.

• After clicking the Restart Button in vCenter Web Client (if available) the task doesn't complete successfully.
• Hosts show as disconnected in the vCenter Web Client.

Environment

• VMware vSphere 7.x
• VMware vSphere 8.x
• VMware vSphere 9.x

Cause

• During the vSAN Cluster Shutdown workflow, advanced parameters are set to 1 to protect data. Specifically, /VSAN/DOMPauseAllCCPs pauses all vSAN component operations. If the restart script fails to run (e.g., due NTP or host connectivity issues, etc), these values are not reverted to 0, leaving objects inaccessible.
  
  
• Running the command "esxcfg-advcfg -g /VSAN/DOMPauseAllCCPs" on all the ESXi hosts confirmes that the value of "/VSAN/DOMPauseAllCCPs" is set to 1. 
  
  esxcfg-advcfg -g /VSAN/DOMPauseAllCCPs
Value of DOMPauseAllCCPs is 1
  
  
• Running the command "esxcfg-advcfg -g /VSAN/IgnoreClusterMemberListUpdates" on all the ESXi hosts confirms that the value of "/VSAN/IgnoreClusterMemberListUpdates" is set to 1. 
  
  esxcfg-advcfg -g /VSAN/IgnoreClusterMemberListUpdates
Value of IgnoreClusterMemberListUpdates is 1

Resolution

1. Revert the advanced parameter values back to 0 across all ESXi hosts in the vSAN cluster by running:
   
   
   1. esxcfg-advcfg -s 0 /VSAN/DOMPauseAllCCPs
   2. esxcfg-advcfg -s 0 /VSAN/IgnoreClusterMemberListUpdates
   
   Verify that VMs are now accessible in the inventory. If the issue persists, contact Broadcom Support.

Additional Information

Related links:

Inaccessible objects in vSAN when all components show as Active

All vSAN VMs are inaccessible after a planned cluster shutdown due to inconsistent cluster power status in the vSAN shutdown workflow