Re: How to move root volume to aggr with less disks?

cgeck0000 · ‎2017-07-25

Hi there.

I've setup our first AFF200 system with ONTAP 9.1. I powered it up and went through the Guided Cluster Setup.

Reviewing the nodes afterwards I see the root aggregate is taking 10 of the SSD 3.8TB drives and set them to 53.88GB of usable space out of the 3.49TB of physical size. The root volume is set to 368GB.

So my first question is what's going on here? Why 10 disks? Why has the system set the drives to 53.88GB of usable space? Is this the standard default setup now? Is it still best practices to have a dedicated root aggregate consuming this many drives?

How can I move this to the standard dedicated 3 drives?

Thanks for the help.

NAYABRSK · ‎2017-07-25

Hi,

Thank you for contacting NetApp communities. starting from Cluster mode NetApp has a special feature called Advance disk partitioning (ADP) where all your disks are shared. To explain you more what ontap does is instead of using whole dedicated disk for root volume ontap picks certain amount of size from each disk (In your case is 53.88GB) and deploy all ontap binaries during initial setup. To verify this information check commands below to confirm if your disks are shared.

storage aggregate show -fields uses-shared-disks

storage disk show -container-type shared

If you still want to corss verify log on to your Oncommand System Manager and go to aggregate and look through the disks in the bottom pane, there you can see same disk shared across aggregates (Check with Disk ID or Name) also can see description as shared.

This is by design on entry level NetApp controllers as to utlise maximum of disks.

Thanks,

Nayab

NAYABRSK · ‎2017-07-26

Found this KB, Should give you a good picture about ADP and moving root volume.

https://kb.netapp.com/support/s/article/ka31A0000008gyeQAA/How-to-non-disruptively-create-a-new-root-aggregate-and-have-it-host-the-root-volume-in-clu...

rlawson · ‎2017-08-24

AFF systems use Advanced Drive Partitioning for better storage efficiency. That root aggregate is sitting on partitions, not on whole drives. The remaining portions of the drives are available for use in data aggregates. The total capacity used by the root aggregates is far less this way.

In addition, the partitioning scheme that comes with all HA-pair AFF systems enables the two controllers to share the high IOPs provided by the SSDs. Each system uses a portion of each SSD. That partitioning scheme is called root-data-data. It provides excellent capacity efficiency as well as top performance.