Nutanix Usable Capacity Calculation

  • 19 August 2015
  • 6 replies

Badge +3

I'm relatively new to Nutanix and am trying to get my head around the resulting net usable disk capacity based on the original raw total disk capacity. As I understand it, Nutanix do not implement RAID (at least not in the conventional sense), however the available usable disk capacity diminishes very quickly and I'd like to understand where it all goes!

I've seen the very useful online Nutanix disk calculator tool. While this is a very slick way to establish how much usable capacity will result from a particular disk selection, it does not enlighten one as to how the usable capacity is arrived at. (see: http://

One question I have is does the Resiliency Factor (RF) account for the N+1 redundancy rule whereby one should always deduct the disk reource of 1 Nutanix Node to allow fot the failure of one Node without any impact on the remaining system. I don't think it does simply because the usable capacity always seems to be a direct multiple of the numbder of nodes selected.

Any help or advice would be gratefully received as I am striving to calculate our disk sizing requirement for a new Nutanix deployment.

Kind regards,


6 replies

Userlevel 4
Badge +18
Hi Nick,

Total User available storage capacity is based on the RF factor configured for the cluster.

You can have RF 2/3 in Nutanix cluster but both have their own set of pre-requistics
Taking an example of RF2 which is most commonly used by customers in Nutanix cluster.
If total storage capacity of entire Nutanix cluster is 10TB then total available capacity which user can use is 5TB and we will store 2 copies of the data to provide redundancy.

If you have any more doubts about this topic, feel free to reply on the thread.

Userlevel 4
Badge +18
I just checked the link which you pasted in your question.

There is some SSD space which we reserve for metadata, different read and write caches but they don't change with the increase of SSD disk size.

If we go back to the site which you mentioned in your post, there are 2 categories Extent store SSD and Extent store HDD. They represent the actual space on which user data is finally stored. But again you have keep RF2 in mind and actual usable space will be half of extent store space (HDD+SSD).

Userlevel 2
Badge +14
If you're considering EC-X Josh Odger's has a nice blog post on cluster scale and usable capacity:
Userlevel 2
Badge +10
you are correct, Replication factor does not callculate for N+1.
  • 4 nodes cluster
  • 16TB RAW
  • with RF2 is approx 8TB usable
  • with N+1 you have to reserve capacity for one node, which is in my expample 2TB
  • this gives you 6TB usable with RF2 and N+1 - you can put host in mainatenance, remove it from cluster and it does not affect cluster availability
let me know if you have more questions
Badge +4
So to acheive RF2 and N+1 with a four node block you can only effectively use approximately 1/3 of the physical storage or the data store will be over allocated?
Userlevel 1
Badge +14
my environment:
6 * 3060-G4 NODE
per node : 2*ssd 480G, 4* HDD(how to find out if it`s sas or sata drive ?)1TB
how to calculate?
I see prism report 22.18TiB MaxCapacity(Physical)。but I think physical should be 1T*4*6+480G*2*6=24T+5760G
if RF2 need two complete copy(like RAID1) ?!, the usable capacity is just half of 24T+5760G, and if after one node failure, for N+1 , I should also reserve another extra node capacity(4T+480G*2) ?!
also I have a basic question and guess, if I have three NODE all, and one node failed, data be replicate to another node because RF, so I have two copy again. but if the second node also fails later, will the cluster be down ? data lost ? I think data shouldn`t be lost because two copy exists.

the customer challenge me for that too, it seems wastes too much space.

any response would be appriciated!