Can use root aggregate to store user data

Jimesh_Makawana · ‎2016-02-29

Hi,

Recently I procure NetApp FAS2552 with 1.2TB 10k X 24 Disk.

By default it's created one aggregate name aggr0 in both node and it's consume 6 disks (Node 1 - 3 disk with RAID DP, Node 2 - 3 disk with RAID DP) with 7 mode.

I search in NetApp community & Library, They said that "Do not store user data in the root volume."

If I create new aggregate with RAID DP in both node, It's consume 4 disk for parity and i get 13 TB usable space

If I add disk to root aggregate I get 17 TB usable space.

Can i use root aggregate to store user data.??

If I store data, Is it create some issue in future? Like create cluster or in expansion of storage.

What is disadvantages to use root aggregate to store user data.

aborzenkov · ‎2016-02-29

Yes, you can. With small number of disks that is the only practical configuration. Even C-Mode now allows it (indirectly via disk partitions, but it is still the same - disks are shared between root and user voumes).

Jimesh_Makawana · ‎2016-02-29

Hi aborzenkov,

Thanks for reply,

We are using 7-Mode, So we can't use Advanced Drive Partitioning.

Can we add disk to root aggregate, Reduce root volume size to 250 GB, Create user volume and and map those volume to user??

Is this is advisable. Is it best configuration.

I am asking this because we plan to create NFS share and map it to unix box which runs SAP. If any configuration issue found later it may down our production servers.

Regards,

Jimesh Makawana

JGPSHNTAP · ‎2016-02-29

Reduce vol0 to 500g

Show us

aggr status -r

Jimesh_Makawana · ‎2016-03-01

Hi,

Thanks for response.

My simple questions is "What is disadvantages to use root aggregate to store user data.?"

Regards,

Jimesh Makawana

JGPSHNTAP · ‎2016-03-01

With that small of a filer, why would you have a data and os aggr.

You need to combine them.. In 7-mode there isn't an issue unless you tax the system.

aborzenkov · ‎2016-03-01

Advantage - disks are shared between multiple workloads.

Disadvantage - disks are shared between multiple workloads.

Seriouosly, there is no black and white, espicially in IT. What is advantage to one, is disadvantage to other. Separate data aggregates are more flexible (you can move data to another filer) and reduce impact of workloads on each other. Multiple small aggregates waste space and reduce maximum performance for each workload. Pick your choice.

NigelM · ‎2016-03-01

Hi Jimesh

in cDOT you cannnot store data in the root aggregate however, as you are running in 7 mode your root aggregate contains vol0 whcih is where ONTAP system files live. It is fine to create volumes in the same aggregate vol0 lives in.

Personally i would want vol0 to be in an aggregate of its own and named accordingly (aggr0_root) which is sized approriatley, but thats just me!

Nigel