#Commands to partition 24x15.3TB SSD's

1 messages · Page 1 of 1 (latest)

bitter rain
#

We have physically added 1 full shelf of SSD's onto 2-nodes HA. Then I assigned the ownership by using commands "storage disk assign -disklist 0-11/12-23 -node node2/node1".

Currently, there is rg0 in each aggregate/node. I am thinking to create a new rg1 and it should consist of 23-new-disks.P1 to the aggr on node2, and 23-new-disks.P2 to aggr on node1.

Next, I would like to partition disks to P1,P2,P3, the same format/size as those existing ones. For that, I remember I need to run a command basically to copy format/size from a existing disk to those new disks. Can you please advice what the command is?

Also, after the partition, I need to assign the ownership of P1 and P2 to two respective nodes, ex, P1->node2, and P2->node1. Correct?

tawny drift
bitter rain
#

@tawny drift Glad to see you here.
After partitions, I did the simulate command as shown in the attachment. Questions:

  1. 13.2.0 and 13.2.1 as shown with no size, is this because the raidtype is raid_dp, same as the existing rg0? I didn't specify riadtype here.
  2. The total usable in the end, 130.7TB, is this as much as we can get?
  3. Can I go ahead to run the command without "simulate"? Just verify if it should be okay.
tawny drift
#

the - - i believe are the parity disks. so not counting towards useable.

#

as far as best... let me check

#

yeah, that's about right. new RG with 6.xTB partitions.

#

i missed the last part of you question "Also, after the partition, I need to assign the ownership of P1 and P2 to two respective nodes, ex, P1->node2, and P2->node1. Correct? " yes. they need to be assigned to the approrate nodes.

#

looks liek they're assigned, but here's an example how to assign partitions.

#

set d <yes> storage disk option modify -node * -autoassign off disk removeowner -disk x.x.x -data true disk assign -disk x.x.x -data true -owner NODEx disk show -fields data-owner, root-owner storage disk option modify -node * -autoassign on set admin

#

the -data could be -data1 - data2. and don't forget to assign the root partitions as well.

bitter rain
#

OK. But, based on the output of "simulate" command as attached, once I formally execute the command(without "simulate"), the ownership should be also assigned and I don't need to use your commands to do any assignment. Right?

tawny drift
#

if it's grabbing them in there, they should be assigned.

#

what's "storage disk show -partition-ownership" show?

bitter rain
tawny drift
#

yep!

bitter rain
#

Good to run the command without "simulate" then?

tawny drift
#

i'd think so from based on what i've seen.

bitter rain
#

@tawny drift One more question I would like to know before execute. By using the same disks(partitions), If I am going to create a new aggr instead of a new raidgroup and adding it into an existing aggr(161TB), both methods will give me the same capacity(130.7TB), the existing aggr and the new aggr to be could share the same existing spare disks(partitions), I believe. Please find the attached "storage aggregate create..simulate" command as attached.

What would be the differences between these two methods, which one is preferred and why?

bitter rain
tawny drift
#

since it's a new RG, you won't gain any more space.

If they're like for like, e.g 15TB drives, same size partitions etc, yes they'll share. Here's a KB in how ONTAP picks drives for failed replacements.
https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_does_Data_ONTAP_select_spares_for_aggregate_creation%2C_aggregate_addition_and_failed_disk_replacement

As far as which to do.. I don't think there's a better way here or one way you'll get more space.

I will say thought that generally, 1 aggr per node is probably easier to manage.

NetApp Knowledge Base
How does ONTAP select spares for aggregate creation, aggregate addi...

This KB article details the disk attributes used by Data ONTAP during spares selection for - aggregate creation, aggregate addition and failed disk replacement.

bitter rain
#

In terms of performance wise, a large aggr (291TB) and two smaller ones(161TB and 130TB), there shouldn't be any difference between these two cases, because they are SSD aggrs, not SPIN aggrs. Right? I promise, this would be my last question.

whole oyster
#

Correct

bitter rain
#

Thank you all.!

tawny drift
#

Bit late to reply. But What Scott said 😁