What happen if we turn on dedup and compression?

  • 2 December 2014
  • 6 replies

Badge +3
Hi, I know it's covered in NPP, but I cannot find the specific line about this one again.
So what happen when we turn on dedup and compression?

This topic has been closed for comments

6 replies

Userlevel 4
Badge +18
When you turn on compression, data is compressed and helps in actual redudction in disk space. There are 2 types of compression i.e
1. Post-process compression 'Data is compressed after it is written'.
This type of compression is recommended for most workloads.
2. Inline compression 'Data is compressed as it is written'.
This type of compression is recommended for workloads that perform batch processing.

Deduplication allows the sharing of guest VM data on Nutanix storage. Redundant VM data is removed and 1 copy of data is kept. There are 2 types of deduplication depending upon whether it is happening on SSD drves (performance tier) or SATA drives (Capacity tier)

Hope it answers your questions.
Badge +3
Thanks for the reply,
what I meant is, will it run together or you can only run dedup but not compression or vice versa
Userlevel 4
Badge +18
Yes they can work together and you can enable both these features in the same Nutanix cluster
Badge +6
Strange, I asked this specific question through a nutanix support ticket and this is the answer I got:

"Regarding dedup and compression please note they do not work together when active in the same container so if your customer has activated them on different containers it should be no problem. Nevertheless if he has to disable dedup this can be done without impact."

Can you please clarify what the official statement by Nutanix is?
Userlevel 4
Badge +18
I am sorry but I just cross checked it and found that when compression is enabled, deduplication will not take effect. You can refer to following document


Check Deduplication best practices. There is a table where it states when not to enable deduplication.

Once again sorry for the confusion.
Userlevel 1
Badge +6
Hi Mathias,

Both statements are correct. The nuance is in the fact that both can be enabled at the same time in a cluster, but they won't currently both be active on a single container. However you can have multiple containers in a cluster, for specific workloads and set compression and dedup when appropriate for that workload.

Just turning both options doesn't mean they will both be suitable.

For databases compression makes much more sense than deduplication, and deduplication is something that's much more effective on for instance statefull VDI workloads. File servers might be better of with post-process compression. However trying to compress a fileserver full op.jpg's or mpg video's might also not gain you anything. So look at the workload and decide what's best for you.

So once again, you can set different dedup and compression policies per container depending on the workload you place inside them. Policies are late-binding which means no hardware configuration during installation is needed, no dependency on special cards, and policies can be modified (or even completely removed) afterwards without destroying any data.

Hope this answer helps!