De-Duplication is it done to all data on the array or individually per lun.

[WPISSystems]

I've been researching and trying to find out if the deduplication on a AF series storage appliance is at the lun level or array level. I see that I have the option to enable or disable at the lun level. but I'm looking for more detailed information on the actual dedupe process as opposed on how to enable or disable.

here is my scenario.

I have a production LUN with data on it.

I have a second LUN with the exact same data on it.

if I enable deduplication on both, does the array dedupe the data that I only consume the storage 1 time.

in this scenario I am unable to use a ZCC in this situation because of reasons to complication to explain in this post. so I'm hoping that I can gain similar space savings through dedupe.

[giladzzz]

Hi

if you are talking about 3PAR it is done at the CPG level and the block size is 16K

if you are talking about Nimble it is done at the array level.

Regards

[Nick_Dyer]

Hello

Nimble deduplication is performed at the POOL level (not array or group), and each application profile you use creates it's own deduplication domain - ie all volumes with the VMware policy will dedupe together at 4k blocks, all volumes with SQL will dedupe at 8K blocks, so on and so on. This is a major benefit of the variable block filesystem of NimbleOS.

Dedupe is enabled globally by default (under Administration->Space). You can disable this, and turn it on/off at the profile or per-volume level if you wish, but I wouldn't recommend it, as it will introduce additional complexity for very little benefit.