You see inaccessible virtual objects while vSAN Skyline Health reports "vSAN Cluster Partition" and "vSAN Object Health" errors.

Products

• VMware vSAN
• VMware vSAN 7.x
• VMware vSAN 8.x

Issue/Introduction

• When checking Virtual Objects (Cluster -> Monitor -> vSAN -> Virtual Objects), you see inaccessible objects. 
• You see a vSAN Skyline Health reporting a vSAN Cluster Partition:

 

• You also see vSAN Skyline Health reporting "vSAN Object Health Error":

 

• All VM's appear to be accessible and running, as expected. 

 

Environment

• VMware vSAN 8.x

Cause

• The partitioned vSAN node was in that state while VM's  were deleted from inventory, so had not updated the Cluster Monitoring, Membership, and Directory Service (CMMDS) properly. 
  
  
• You can validate the owning host of the inaccessible objects from an ssh session on one of the ESXi hosts in that cluster as follows:

esxcli vsan debug object list -u <uuid>

...where the UUID of the inaccessible object is found in the Virtual Object List (Cluster -> Monitor Tab -> Virtual Objects -> Click on "Inaccessible")

NOTE: When you perform the same command on one of the hosts that is still in the primary partition, the object is not found. 

 

Resolution

• You will need to troubleshoot and resolve the vSAN Cluster partition first. For more information on troubleshooting vSAN Cluster Partitions, see KB: vSAN Health Service - Network Health - vSAN Cluster Partition.
  
  
• If the host appears to be no longer partitioned (e.g. all nodes can ping each other, all firewall ports are confirmed open, etc.), but a "retest" in Skyline Health (Cluster -> Monitor Tab -> Skyline Health -> Retest) still reports the host in a partitioned state, place the host into maintenance mode and reboot it. 
  
  
• If you are having trouble resolving the partition or if the partitioned state is cleared and you still have inaccessible objects after the reboot, please open a support case with Broadcom Support for further assistance.