VMware Cloud Community
RomanoR
Contributor
Contributor
Jump to solution

provisiond space of datastore doesn't match with space of the VMs

Hi guys,

I have special phenomenon on my cVenter environment. The provisioned space on the datastores doesn't match with the really used capacity of the VMs.

Details:
Vcenter: VMware vCenter 6.0.0, 3634793
ESXi: ESXi 5.5.0, 3568722
Storage: Netapp Cluster Mode
NFS datastores

The VMs allocate more space as assigned. It could be 100% more space. In the datastore there are no files which are responsible for this. On the Netapp we have enabled Storage Efficiency which means deduplication.
From the vCenter view - the datastore has a provisiond space of about 1TB and really space of about 550GB (Space of every single VM).

Do you  know this issue or something like that??

Best regards

Reply
0 Kudos
1 Solution

Accepted Solutions
RomanoR
Contributor
Contributor
Jump to solution

Thanks for your answers but it is a Netapp feature called fractional reserve.

This feature reserved 0% or 100% of space.

Has someone experiences with the fractional reserve option and NFS volumes for ESXi?? What is the best value for this?

Romano

View solution in original post

Reply
0 Kudos
8 Replies
Nick_Andreev
Expert
Expert
Jump to solution

If you use thin-provisioned .vmdk disks then the amount of provisioned space will be equal to the sum of the .vmdk disk sizes, but amount of space used will be equal to the amount of space actually allocated to VM disks, which will be smaller.

Hope that makes sense.

---
If you found my answers helpful please consider marking them as helpful or correct.
VCIX-DCV, VCIX-NV, VCAP-CMA | vExpert '16, '17, '18
Blog: http://niktips.wordpress.com | Twitter: @nick_andreev_au
Reply
0 Kudos
RomanoR
Contributor
Contributor
Jump to solution

Thanks for your answer. But the VMs used lazy-thick or eager-thick provisioned hard disks.

Nevertheless, the space of the vms and the datastore don't match.

Reply
0 Kudos
Rubeck
Virtuoso
Virtuoso
Jump to solution

Nevertheless, the space of the vms and the datastore don't match.

I'm guessing that this is due to vCenter not knowing about the dedup done by the NetApp.... So seeing 4x 500GB VMs on a single 1TB datastore (just an example) would be expected.   

/Rubeck

Reply
0 Kudos
HawkieMan
Enthusiast
Enthusiast
Jump to solution

You should also take into consideration that nvram and swap is not calculated into the used storage space. As well as snapshot reserves.

Reply
0 Kudos
RomanoR
Contributor
Contributor
Jump to solution

Yes, I know.

But my issue is that we have 4x50GB VMs and in the datastore are 500GB allocated.

Reply
0 Kudos
Nick_Andreev
Expert
Expert
Jump to solution

Run RVTools against your vCenter. It performs some health check tests, which can show you if there are any "zombie" files on your datastores.

---
If you found my answers helpful please consider marking them as helpful or correct.
VCIX-DCV, VCIX-NV, VCAP-CMA | vExpert '16, '17, '18
Blog: http://niktips.wordpress.com | Twitter: @nick_andreev_au
Reply
0 Kudos
RomanoR
Contributor
Contributor
Jump to solution

Thanks for your answers but it is a Netapp feature called fractional reserve.

This feature reserved 0% or 100% of space.

Has someone experiences with the fractional reserve option and NFS volumes for ESXi?? What is the best value for this?

Romano

Reply
0 Kudos
Nick_Andreev
Expert
Expert
Jump to solution

The idea behind the Fractional Reserve is to reserve the amount of space equal to the volume size for the snapshots. Very controversial feature, I tend to disable it and just monitor the space on the volume to make sure snapshots do not use too much space.

And if you're not using snapshots on the NFS volumes, then disable it without a doubt as you're just wasting your space.

---
If you found my answers helpful please consider marking them as helpful or correct.
VCIX-DCV, VCIX-NV, VCAP-CMA | vExpert '16, '17, '18
Blog: http://niktips.wordpress.com | Twitter: @nick_andreev_au
Reply
0 Kudos