VMware Cloud Community
pcrane213
Contributor
Contributor

ESXi 6.7.0 - Space Reclaimation HELP!

Hello,

We have been searching for hours and following steps on how to reclaim space on a thin provisioned VM with no luck.

Nothing seems to be working.

We have tried - https://community.webcore.cloud/tutorials/how_to_reclaim_disk_space_from_a_thin_provisioned_/

Nothing changed afterwards.

And UNMAP command is not supported.

These are some outputs

VAAI Plugin Name:

   ATS Status: unsupported

   Clone Status: unsupported

   Zero Status: unsupported

   Delete Status: unsupported

   Revision: 3.45

   SCSI Level: 5

   Is Pseudo: false

   Status: on

   Is RDM Capable: true

   Is Local: true

   Is Removable: false

   Is SSD: false

   Is VVOL PE: false

   Is Offline: false

   Is Perennially Reserved: false

   Queue Full Sample Size: 0

   Queue Full Threshold: 0

   Thin Provisioning Status: unknown

   Attached Filters:

   VAAI Status: unsupported

Reply
0 Kudos
14 Replies
pcrane213
Contributor
Contributor

To clarify. The VM is thin provisioned for 200GB and current used space is only at 30GB but the VMDK is 199GB.

Reply
0 Kudos
blazilla
Enthusiast
Enthusiast

Can you provide more information about the underlaying hardware?

Best regards Patrick https://www.vcloudnine.de
Reply
0 Kudos
pcrane213
Contributor
Contributor

Manufacturer

Supermicro

Model

Super Server

CPU

8 CPUs x Intel(R) Xeon(R) CPU D-1540 @ 2.00GHz

Memory

127.9 GB

Reply
0 Kudos
blazilla
Enthusiast
Enthusiast

What about disk configuration, VMFS etc?

Best regards Patrick https://www.vcloudnine.de
Reply
0 Kudos
pcrane213
Contributor
Contributor

Datastores are Non SSD - VMFS5

Reply
0 Kudos
blazilla
Enthusiast
Enthusiast

Looks like your underlaying storage does not support VAAI. What kind of storage do you use?

Best regards Patrick https://www.vcloudnine.de
Reply
0 Kudos
RickVerstegen
Expert
Expert

What guest operating system is the vm running?

Typically, the unmap command that generates from the guest operation system on VMFS5 cannot be passed directly to the array. You must run the esxcli storage vmfs unmap command to trigger unmaps for the array.

However, for a limited number of the guest operating systems, VMFS5 supports the automatic space reclamation requests.

To send the unmap requests from the guest operating system to the array, the virtual machine must meet the following prerequisites:

1. The virtual disk must be thin-provisioned.

2. Virtual machine hardware must be of version 11 (ESXi 6.0) or later.

3. The advanced setting EnableBlockDelete must be set to 1.

4. The guest operating system must be able to identify the virtual disk as thin.

Was I helpful? Give a kudo for appreciation!
Blog: https://rickverstegen84.wordpress.com/
Twitter: https://twitter.com/verstegenrick
Reply
0 Kudos
pcrane213
Contributor
Contributor

Not sure how to check this, would it be Local Storage?

Reply
0 Kudos
pcrane213
Contributor
Contributor

Running Windows Server 2012 R2.

We have tried to follow the UNMAP command through SSH to the host but it states it is not supported.

Reply
0 Kudos
pcrane213
Contributor
Contributor

Here is a screenshot - Awesome Screenshot

It states 163.13GB free but Used is 200GB? so the VMDF is 199GB~

Reply
0 Kudos
Mukz9
Enthusiast
Enthusiast

Hi

The Automatic UNMAP/Storage space reclamation may not be supported on your storage array. You can determine whether the storage array supports automatic unmap/storage space reclamation by referring to the VMware HCL. The blog https://cormachogan.com/2016/12/05/determining-array-supports-automated-unmap-vsphere-6-5/  provides detailed procedure to determine if the storage array supports automated UNMAP of unused storage space.

Reply
0 Kudos
blazilla
Enthusiast
Enthusiast

I'm sorry, but it looks like your setup doesn't fulfill the requirements to support UNMAP/ Space Reclamation. You can try to zero-out the guest filesystem with sdelete.exe and move the VM onto an NFS datastore and back to your local storage. Otherwise you can shutdown the VM after the zeorout operation and use vmkfstools --punchzero to reclaim the space.

Best regards Patrick https://www.vcloudnine.de
Reply
0 Kudos
pcrane213
Contributor
Contributor

Thank you again for trying to assist everyone.

We have tried to zero-out the guest filesystem with sdelete.exe and then shut it down then vmkfstools -K [disk].vmdk.

We assume because VAAI Status is unsupported and Delete Status is unsupported it didn't work. Nothing was changed.

What are our options to reduce this vmdk?

When you say moving the VM onto a NFS datastore and back to local storage could you confirm the exact steps for this. Also, would it work? Is there any other way?

Could we simply unregister the VM then register it?

How does moving the folder to NFS then moving it back and registering it be different to unregister it then just simply register it again?

Reply
0 Kudos
pcrane213
Contributor
Contributor

Hello,

Could anyone please assist?

Reply
0 Kudos