Dell Compellent vs VMware -> incorrect free-used space report

You told us that Compression and Dedup is enabled on that volume? Whats the space saving ratio?

Notes:

• I have seen the same earlier that SCOS reportet way less storage usage as Hosts
• IIRC in one of the latest release notes they mentioned a fix for wrong free space calculation within DSM/SCOS, but your 7.4.x sounds already like the latest version

Regards,
Joerg

Hello 

Today the picture is little bit different, RATIO decreased from 3.8 to 2.8 (how?) but saved space is insane!

vCenter continue to show me the same 11TB in use

4.12TB * 2.78 = ~11.4TB

I think its pretty close.

It is close and (let's presume) this is correct.

The question is how to reflect 4.12TB used in vmware ?

This is not possible because

1. Its Block storage and not NFS where there storage server have control through the filesystem
2. SCOS is thinproivsioning by default
3. Depends on several things you are effected by not having an scsi unmap. It means that when your ESXi delete files (Snapshot deleting, removing VMDK/VMs) on VMFS your no one tells* this to your SC volume. So in a normal VMware environment the space usage on SC is higher than on your Servers (VMFS Datastore). You have to run scsi unmap on your datastores from time to time or using VMFS 6.0 and a <= 1MB Blocksize on SCOS side (there are some more requirements for the automated scsi unmap!)
4. Your SC uses space savings like Compression/Dedup which it totally independent from the server perspective

* Think about when deleting Files on Windows.  The OS marks this in the File allocation table but not phy. delete the files. Similar is on a Block storage when the server doesnt tell the storage to mark blocks as deleted.

Regards,
Joerg

Regarding "it is not possible" - why it is possible in NetApp world?

https://www.flackbox.com/netapp-lun-space-reclamation

No problem to run UNMAP command and it is planned for this weekend. I hope this will actually do the magic 

Well... you dont get it.   I have read the URL you posted and its even more worse (just a little) on a Netapp because you have to enable space reclamation on the lun first and than you need a supportet OS running on the hosts.

On a SC you only need a supportet OS or some help on the server side...... well and a supportet environment.

What the URL not mentioned is the use of Compression/Dedup.  So when you would like to see matched space usage from both perspectives than you have to disable Compression/Dedup and be sure you have a Server OS with automatic space reclamation or just run it periodically. Than you will see the same behavior as with a Netapp. 

Take a look to your Tier1 usage. The SC have take notice of 11.42TB saved data and the data reduction crunch it down to 4.63TB which is a great value by the way. From the numbers i assume that there was not a lot of deleted data on that volume.

Regards,
Joerg