topic Re: VSA 4TB limit question in StoreVirtual Storage

VSA 4TB limit question

Mike Tree — Mon, 27 Nov 2017 11:45:46 GMT

Hi,

I wondering if anyone can help, we are looking like to purchase this:

HPE StoreVirtual VSA 2014 Software 4TB 3 Pack E-LTU Q5T60AAE

I believe it allows us to install the software on 3 x 4TB raw storage nodes, my question is will it allow us to create 6TB worth of RAID10 luns split across the 3 x nodes or 8TB RAID5 luns split across the 3 x nodes?

thank you in advance

Mike.

Re: VSA 4TB limit question

oikjn — Mon, 27 Nov 2017 16:09:36 GMT

for NR10, your consumed space is 2x the usable space, so with 3x4TB nodes, you have roughly 12TB of raw space which will yield you 6TB of usable space when configured as NR10.

Re: VSA 4TB limit question

Mike Tree — Mon, 27 Nov 2017 16:13:18 GMT

Hi oikjn,

thanks for your reply, will the license allow us to create 6TB's of Volumes? or will it limit the volumes to 4TB ?

thanks

mike

Re: VSA 4TB limit question

Fran_delaTorre — Wed, 25 Apr 2018 08:46:53 GMT

Hello Mike,

4TB Storevirtual liceses allow you to create 4TB VSA storage nodes then once the Mangaement group is created and then the cluster you will be able to use all cluster capacity as you wish creating volume of any size.

Notice what is previously commented that Network RAID 10 will consumre 2x volume size space.

Re: VSA 4TB limit question

ThanhDatVo — Fri, 27 Apr 2018 08:06:27 GMT

Hello Mike,

You can create as many LUN's on the shared cluster (12TB raw space) as you'd like with any size as you'd like.
Each LUN can be defined as Network RAID 10 or Network RAID 5. You can also mix NR10 LUN's and NR5 LUN's on the same cluster
For each LUN you can choose between thin or thick provisioning (with NR5, only thin provisioned is possible, NR10 supports both types)

You can store a theoretical max of 6TB in NR10 or 8TB of NR5. Keep in mind that you do need some free space on each LUN & the cluster for your storage to perform properly.
If you need performance, NR5 is definitely not recommended. Also performance degradation is insane when one node goes down.
A single LUN for the entire cluster is also configurable but not recommended.

Multiple LUN's with a thick provisioned NR10 is recommended.

Kind regards,

Thanh Dat Vo