Space Usage Not Matching

[Macpilz]

Good Day

I have a weird issue, My host is saying I am using more space than what's shown inside the VM. I have 3 VMDKs connected to the same host and these VMDKs are from a single SAN. The sizes of each disk from the SAN are 9TB, 16TB and 17TB, usage according to VMWare is 7TB, 12TB and 15TB respectively. Account to the host usage is 5TB, 6TB and 12TB respectively.

We currently use Veeam for backups so suspect it created some sort snapshot that I just can't find, I've consolidated the disks to try and recover that space browsed on the VMDKs to see if I can find but there are no delta files.

Not sure where to start with this so any help would be appreciated

Thanks

[ashilkrishnan]

Good day

Please share some screenshots for reference and check the following as well

1. Are the VM disks thick or thin provisioned ?

2. Is the datastore usage different in vCenter and directly connected to one of the ESXi host ?

3. Browse the datastore and check if it has any other files apart from running VMs ?

4. Please SSH to one of the ESXi hosts and run following commands to check the usage:

df -h

du -sh

[Macpilz]

The VMs are thin provisioned and the datastore usage is the same on both the vCentre and ESXi. There is nothing but the VMDK and a ctk file.

I've attached a picture of the 3 volumes in question. The other command just shows the total size of the disks attached the host and I have no issues with the rest of the disks, just these 3.

[ashilkrishnan]

Thank you for sharing these details

Can you share some screenshots where it shows the difference of VM usage and datastore usage?

[Macpilz]

Good Day

These are the drives in question, I haven't fully used the available storage from the VMFS volume but the data usage doesn't match with what is used on the drive.

[ashilkrishnan]

Thank you for sharing these details

Please check the size of .vswp files in VM folder within the datastore. This would take up additional space apart from the space utilized by VM disk files

If disk space was recently cleared in the VM, you can refer space reclamation document -

Space Reclamation Requests from VMFS Datastores

[Macpilz]

Where do I see the .vswp files, nothing like that in the Datastore. The disks are VMFS 6, doesn't that automatically make them reclaim the space?

[ashilkrishnan]

Yes, VMFS6 will reclaim space automatically depending on the type of reclaim setting. vswp files can be found inside the VM folders

[Macpilz]

Searched for it in the right location and found the swap files you were referring to, that's a totally different datastore, these datastores are primarily for storage, no RAM will be linked there.

[ashilkrishnan]

Alright. Comparing the output from df -h and the datastore usage summary from Web client, they seem to be matching. Please compare the following:

VM storage usage on VM summary

Navigate to Datastore/VM folder from SSH and run du -sh to list the actual usage

[Macpilz]

Hi Ashil

That command only listed the disk usage, what I have noticed is there are snapshots not being deleted when I run deletion option, going to reboot the server as a troubleshoot option and then run updates.

Will give you feedback tomorrow.

[ashilkrishnan]

Thank you for the update

You can perform the following to check if the VM is running on snapshots when snapshot manager does not list any snapshots:

Go to edit settings of the VM

Expand the hard disks and check the disk where it is pointing to: Regular vmdk: VMname.vmdk and snapshot disk: VMname-00001.vmdk

If the VM disks are not pointing to a snapshot disk, you can such snapshot delta disks out of the datastore to another datastore for now and delete it later

If the VM is pointing to a snapshot disk and snapshot manager does not list any snapshots, you can create a new snapshot and initiate delete all.

[Macpilz]

Tried doing updates and the space was not reclaimed. I have backups running everyday from Veeam and snapshots are created and deleted everytime without reclaiming that old space.

one of the disks has a sesparse vmdk inside the list of files but these only add up to about 200GB.

Any other solutions you'd suggest?