Critical errors indicating inaccessible objects and "Stats DB object conflicts" and unable to delete the objects under vSAN Skyline Health

Products

• VMware vSAN

Issue/Introduction

The following symptoms are observed in vSAN Skyline Health:

Inaccessible objects reported under Data Health – vSAN Object Health

Multiple statsdb objects reported under Performance Service - Stats DB object conflicts check

Observing the details of the inaccessible objects in the vSphere Client under Cluster - Monitor - Virtual Objects - Inaccessible Object - View Placement Details reflects only RAID0 object components.  No RAID1, RAID5, or RAID6 object components are visible.

 

Environment

8.0 U3

Cause

The duplicate statsdb and inaccessible objects are a result of using RAID0 storage policies.  Since the objects have no redundancy, the loss of any single vSAN disk, disk group, or host will result in inaccessible objects.  An inaccessible statsdb file will result in the vSAN performance service failing, the database is recreated when the service is restarted.

Resolution

Delete stale statsdb files per KB 317847 instructions, any inaccessible vSAN objects as a result of the use of RAID0 storage policy should be restored from backup as the only other recovery possible is to bring back previously failed hosts or disks, which are often permanently lost.

RAID0 storage policies are not recommended, change the default storage policy to reflect a fault tolerant storage policy, reference storage policy documentation for a full description of policy fault tolerance and rules:

Using vSAN Policies

vSAN objects can be deleted through the vSphere Client datastore browser, if that fails open a case Broadcom vSAN support