ONTAP Hardware

6x SSD for new shared aggregation - is possible?

Jonis
2,755 Views

Hi!

We have a FAS2750 with 12x 960Gb SAS Disks and one aggregation shared between both controller nodes (half of the disks in one controller, other half in second controller, buth only 1 total aggregation).

 

We just brought 6x new 1.6Tb SSD's and I'm trying to create a new aggregation, but without success. Actually, almost. I'm using ONTAP GUI interface to create this new aggregation. Durin wizard, it say that is not possible to automatically determine the correct aggregation because we have less disks than what is needed. Ok.

Them, I switch to manually create and proced on wizard. After all, I have onde aggregation with 5x disks + 1x spare. The problem is that all these disks are owned by only one controller node, not 'shared' like the other aggregation.

So my question is: is possible to create a new aggregation type 'shared' with only 6x disks? (We don't want to mix SSD's and non-SSD's since we already have non-SSD's in production).

 

You can see the difference here in this screenshot:

netapp.png

 

1 ACCEPTED SOLUTION

Fabian1993
2,610 Views

Aggr0 is running on the Root partitions, If you assign your 6 SSDs to Controller A and Build a 5+1 Aggregarte on the Home Node Controller  A, in a failover the Controller B will temporialy the Owner of that Aggregate.

 

 

To assign a Disk to both Nodes you need ADPv2, AFF oder FAS System with SSDs..

What ADPv2 does is to partition the Disk into 3 Partitions, one root, two data - each Data Partition will be assigned to another Node.

 

 

View solution in original post

5 REPLIES 5

Fabian1993
2,736 Views

Shared means that you use ADP, in this Case you use ADPv1 means Root-Data Partitioning,it is possible to create via CLI manually Partitions.

 

to create two Data Partitions on a Disk..


For example, for a root aggregate layout of 8D+2P+2S and the root aggregate size of “112958795 blks”

  • Divide the total number of 4K blocks equally amongst all data drives:

(112958795 / 8 = 14119849.375
ROUNDUP(14119849.374) = 14119850

 

cluster::> run local
Type 'exit' or 'Ctrl-D' to return to the CLI
cluster-01> priv set diag
Warning: These diagnostic commands are for use by NetApp
         personnel only.
cluster-01*> disk partition -n 3 -i 3 -b 14119850 2c.11.3 
disk partition: Disk 2c.11.3 is successfully partitioned.

 

Jonis
2,613 Views

Root partition isn't for operating system? IF so, I don't need another one....I'll keep my current setup, just need to add these 6x SSD's the way that they can work on both controllers (failover) automatically.

 

Fabian1993
2,611 Views

Aggr0 is running on the Root partitions, If you assign your 6 SSDs to Controller A and Build a 5+1 Aggregarte on the Home Node Controller  A, in a failover the Controller B will temporialy the Owner of that Aggregate.

 

 

To assign a Disk to both Nodes you need ADPv2, AFF oder FAS System with SSDs..

What ADPv2 does is to partition the Disk into 3 Partitions, one root, two data - each Data Partition will be assigned to another Node.

 

 

Jonis
2,604 Views

 


@Fabian1993 wrote:

Aggr0 is running on the Root partitions, If you assign your 6 SSDs to Controller A and Build a 5+1 Aggregarte on the Home Node Controller  A, in a failover the Controller B will temporialy the Owner of that Aggregate.


 

This can apply to controller B too? I mean, now I have these disks (SSD's) owned by controller B and one aggregate created. In case of failover, the controller A will owner this aggregate (temporarialy), like you said that controller A would do?

I don't really need to 'share' disks like I said, I just need to make sure that if onee controller goes down, the another one take controll of this new aggregation.

 

 

 

Fabian1993
2,600 Views

Exactly - In a failover the HA Partner will temporarily take the Ownership of the failed Partner Aggregate. So you absolutely fine with that Config.

Public