vSAN -- Health Alarm -- "Thick-provisioned VMs on vSAN"

Products

• VMware vSAN

Issue/Introduction

Skyline Health displays an alert for "Thick-provisioned VMs on vSAN"

Example:

 

The overview section within the alert will detail the disks in question.

Example:

 

Environment

VMware vSAN (All versions)

Cause

This alarm indicates that there is at least one virtual machine disk on the vSAN Datastore which is thick-provisioned.

 

In most cases, this is because the object space reservation (OSR) on the vSAN storage policy is set to 100 - which is the equivalent of thick provisioning for vSAN.

 

This may also be the result of a virtual machine being backed up from another type of SAN and then restored to vSAN datastore. In this case, the disk can be seen as thick provisioned, even if the OSR in the vSAN storage policy is set to 0 (thin provisioned).

 

- Regular Storage vMotion will always provision VMs with thin disks.

 

Resolution

Check the vSAN Storage Policy which is assigned to the Virtual Machine disk.

 

• If the Object Space Reservation (OSR) is set to 100 (thick provisioned), edit the policy to change the Object Space Reservation to 0 (thin provisioned).

 

• If the Object Space Reservation (OSR) is set to 0 (thin provisioned) but the above alert still shows for the disk, follow the steps below to resolve this.

 

Change configuration from 'Thick' to 'Thin' by applying a cloned vSAN Storage Policy to the Virtual Machine disk(s) listed in the vSAN Alarm:

 

First, create a new thick provisioned vSAN storage policy, and apply it to the affected disks.

1. Clone VM Storage Policy applied to the Virtual Machine disks listed in the alert.
2. Edit the new policy to have Object Space Reservation of 100.
3. Apply the cloned Policy to the listed Virtual Machine disk(s) in the alert.
   1. Right-click the on VM in the vSphere Web Client navigator and select "Edit Settings"
   2. On the tab "VM Storage Policies", edit the Storage policy.
   3. Select the cloned Policy from the dropdown list and click "Apply to All". 

 

The disks now have a new vSAN storage policy applied to them, with an OSR of 100.

Next, reapply the original storage policy (the one with OSR set to 0) to these disks.

• Apply the original Storage Policy to the listed Virtual Machine disk(s) in the alert.
  1. Right-click the VM in the vSphere Web Client navigator, and select "Edit Settings".
  2. On the tab "VM Storage Policies", edit the Storage policy
  3. Select the original Storage Policy from the dropdown list and click "Apply to All". 

 

Important: This process will recreate the object twice so will trigger IO operations. This should be done individually or in small batches.

 

Note:

• If there are any snapshots present on the VM, the disk provisioning will not change. Remove the snapshot and retry the solution.
• Sometimes, even after changing the provisioning back to thin, the skyline health alarm shows this VMs name. In those cases, restart the vmware-vsan-health service on the vCenter to fix the issue by using the below command:

service-control --restart vmware-vsan-health

Additional Information

• This Health Alarm will not check VMs deployed by ESX Agent Manager (EAM), as most of the VMs deployed are thick-provisioned by default or VMs with disabled tasks (vm.disabledMethod), for example, NSX Controllers

• With Virtual Machine disk(s) configured as "Thick provisioning", vSAN datastore utilization will be higher. See KB 396137 and VMware Blog for additional information.