Enthusiast
Enthusiast

Provisioned Space Not Equal to Actually Provisioned Space

Hey All,

I have an odd issue with some NFS Datastores in my Environment not reporting the correct Provisioned Space on the Summary Tab of the Datastore. I have 2 of my datastores showing the inconsistencies. I am running VSphere 5 U1. When I click on the Virtual Machines Tab, these are my results:

Datastore 1:

- Summary tab showing 945GB provisioned

- When doing the math via SSH, and Virtual Machines Tab, I get 1881.72 actually provisioned, 1479GB Used.

Datastore 2:

- Summary tab showing 509.79GB provisioned

- When doing math via SSH, Virtual Machines Tab, I get 791.58GB actually provisioned, 512.33GB Used

I just want to set a couple of things straight, as I've checked each and every one:

1) There are no snapshots running on any VM's in the Datastores

2) All VMDK's are accounted for as belonging within the datastores. There are no VM's with Disks in Multiple Datastores.

3) No ISO's are mounted

4) Not only is the Provisioned Space showing way under what is actually provisioned, but even the actual Used Space is higher than the Provisioned Number that vCenter is pulling from somewhere

5) Refreshes were done before pulling any numbers, VCenter was also rebooted

I've included pictures of Datastore 2. The images are from the Summary Tab, and the Virtual Machines Tab. Many of these VM's were just SVMotioned from FC LUNs to NFS Volumes.

Anyone have any suggestions?

http://vvcap.net/db/UA4oGgQT_ar54FrOORs_.pnghttp://vvcap.net/db/xE3mUEbYFLQp5njt-YU-.png

0 Kudos
4 Replies
Enthusiast
Enthusiast

Figured out the cause. I created a brand new NFS Datastore, and provisioned 4 VM's with 20GB  piece. The Provisioned space read correctly at 80GB Provisioned. I then deduplicated the volume on the Netapp side, and I could see my Provisioned space lowering as dedupe was running. After dedupe finished, my Provisioned space was exactly 19.54GB. All 4 VM's were built off the same template.

This tells me that the VMWare Summary tab is not just looking within the NFS Share for VMDK Files. It must have some sort of integration with the Netapp Filers to modify this number. This seems odd to me, but I guess I'll need to check with Netapp on whether or not this is expected behavior.

If anyone has more insight into this, please post. Thanks.

0 Kudos
Immortal
Immortal

Its not actually itegration with netapp directly...it just relies on the filer to tell it used space, not adding up all the provisioned values.   Any NFS device will dedupe will display this.

--Matt VCDX #52 blog.cowger.us
0 Kudos
Enthusiast
Enthusiast

Shouldn't Provisioned Spade be something VMWare takes its own numbers from? I can understand it getting the Available Space and Total Capacity from the Filer, but it doesnt seem right to pull the Provisioned Space numbers from there. That number basically becomes meaningless to VMWare for alarming purposes.

0 Kudos
Immortal
Immortal

If its pulling available space from the filer, I would argue it should also pull use space too Smiley Happy

But feel free to file it as a bug Smiley Happy

--Matt VCDX #52 blog.cowger.us