ONTAP Discussions

Is it possible to create an aggregate? (RAID group with ADP and non-ADP)disks?

YXMRETK

Hi Everyone

We are currently designing to build the FAS2720 from next month.

When the equipment is delivered, it is an OEM product and is already set up.

The 12 disks of the controller have an ADP configuration, and all but the Root aggregate are in an unbuilt state.
It has a disk configuration of 36 NL-SAS 4TB including controller and shelf.

In this case, is it possible to create an aggregate with a 15D + 2P RAID group using 12 ADP data partitioned disks and 5 unpartitioned disks?

When I checked the website, the 5 discs were partitioned, but it was possible to create them, but is that correct?

The details of the device configuration are described below.

HW: FAS2720
Shelf: DS212C (2shelf)
Disk: NL-SAS 4TB (36disk)

Thanx

1 ACCEPTED SOLUTION

TMACMD

Habit. I want to make sure that ontap will in fact use the adp drives first to make the start of the aggregate. I know that if I add disks to an aggregate with adp ontap will auto partition  as long as I am under the max raid size. 

 

 i think you may be able to just do all 16 at once. Just make sure that you don’t inadvertently end up with two raid groups

 

 on new systems (that don’t have a cluster born date before ontap 9.4) the system will use a fast zero to add drives. Really fast and slick

 (if the cluster was born on 9.3 or earlier, disks will always slow-zero)

 

 

View solution in original post

10 REPLIES 10

YXMRETK

postscript
ONAP: 9.9.1

paul_stejskal

I believe you could create a RG1 of the non-ADP disks on the external shelf.

 

YXMRETK

Sorry at the lack of explanation.

 

The configuration this time is based on the following assumptions.

Node 1
Disc: 17 + 1hotspare

Node 2
disk
Disc: 17 + 1hotspare

On both nodes, I need to create an aggregate using 17 disks (15D + 2P).
Therefore, one node needs to have a mixture of ADP and non-ADP disks ...

paul_stejskal
Ah. That makes sense. Honestly I'm not sure if it is possible to create the external shelf as ADP if it is HDDs, and obviously you don't want to reset the system to lose the 6 disks for 2 root aggregates. https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Systems/FAS_Systems/FAS2720_does_not_support_root_aggregate_partition_on_external_shelf

YXMRETK

I couldn't see the link with my authority.

Would you please tell me what you wrote if possible?

TMACMD

Personally, I’d do this

 

 1. create a data aggregate of 6 on each controller using raid dp with the ADP drives. 

2. set the Max raid size to 16 

3. add 10 drives to each aggregate

 the system will auto partition

 

 you will not lose too much to parity as it is already accounted for

 

 you end up with a single raid group on each node of 14+2, 2 spare disks

YXMRETK

Thank you for your answer.

 

Please tell me the content of your answer.

Why not create aggregates on 16 discs at once?

What is the reason for creating an aggregate with 6 disks first?

paul_stejskal
I think ONTAP will allow to do what you want is what Thomas is saying. Create the two data aggregates then when you try to add the disks ONTAP will be like "oops sorry, you need a new RG".

TMACMD

Habit. I want to make sure that ontap will in fact use the adp drives first to make the start of the aggregate. I know that if I add disks to an aggregate with adp ontap will auto partition  as long as I am under the max raid size. 

 

 i think you may be able to just do all 16 at once. Just make sure that you don’t inadvertently end up with two raid groups

 

 on new systems (that don’t have a cluster born date before ontap 9.4) the system will use a fast zero to add drives. Really fast and slick

 (if the cluster was born on 9.3 or earlier, disks will always slow-zero)

 

 

YXMRETK

habit! I think it's a good habit. I want to apprentice.

 

Thank you for your polite answer!

I was relieved to know that I could do what I wanted to do.

 

I'm not good at English, so I used a translation tool to ask questions, so I think there were some parts that were difficult to understand ...

Public