vSAN online health alarm 'Thick-provisioned VMs on vSAN'

[VeeJee]

Hi,

Is there a way to get an alert in vROPS or vRLI when the vCenter alert vSAN online health alarm 'Thick-provisioned VMs on vSAN' is raised?

I can't find anything about it.

Need some help :smileyhappy:

Cheers!

Vincent

[lucasbernadsky]

Hi, VeeJee. In vSAN there is no such concept as "Thick Provisioning". Insted of that you have the Object Space reservation in the storage policies Understanding VMware Virtual SAN Storage Policies -Detailed View

This is not a big problem, but is not compliant for vSAN policies and best practices. It is probably that those VMs were deployed as thick provisioning, or migrated from another cluster.

You can solve this problem doing a Storage vMotion to another datastore (maybe local datastore if you don't have other option) and migrating it back to vsan and in the migration configuration you set up the storage policies for that VM, and problem solved

[vrm]

Hi Lucasberndsky,

I'am understanding what you are saying. But these disks where thick-provisioned due a failure in vRA blueprint So we do have thick-provisiod disks on vSAN as the vCenter alarm displayed. If a thick provisioned disk is needed we do have a storage police with the Object Space reservation is 100. And yes it's a challange to make the disk thin provisioned if a cluster have only one vSAN datastore and planning downtime for a vm to clone is difficult.

So we're thinking the create a small  temporary cluster in the same vCenter with the same vMotion network. Then it's possible to do a vMotion/storage vMotion to this cluster and back to the vSAN cluster with the correct storage policy.

With the following Powershell one-liner I created an overview of disks that are thick-provioned.

$thick_provisioned_servers = Get-Datastore | Get-VM | Get-HardDisk | Where {$_.storageformat -eq "Thick" } | Select Parent, Name, CapacityGB, storageformat, ConnectionState | Export-Csv "d:\temp\\thick_provision_vmdk.csv" -NoTypeInformation

Thank you for reply!

Vincent

[lucasbernadsky]

Great, Vincent! Yup, the procedure needed to change disk format unfortunately involves another datastore. This is a common problem in vSAN environments where there is only one datastore. Creating a new cluster would be a good option if you have licenses available. If not, you can always add a temporary NFS share and add it to all hosts in cluster. Then doing Storage vMotion and applying the new policies on the fly.

Regards

[Dave Morera]

You don't need to create another cluster to have access to NFS datastores, etc. A vSAN cluster is a vSphere cluster, so you can still have other datastores within the same vSAN cluster and do a storage vMotion within the vSAN cluster.

[VeeJee]

Hi,

I am aware of that. I prefer for a host / cluster with fc storage instead of NFS to the vSAN cluster. But I think we're now off topic :smileyhappy:

Vincent