Disk Config - 2040HA + DS4324

[nsitps1976]

I have a configuration as follows – I want to maximise capacity without sacrificing performance:

1 x 2040HA – 12 x 300GB 15k

1 x DS4342 – 24 x 300GB 15k

I was thinking of splitting the disks 50/50 and having 18 per controller; however, I do not want to have 2 spares per controller – Can I increase RG0 to 17 disks (15 D + 2 P) or is it a bad idea to deviate from the recommended for just 1 disk? – Also, what about 12 for one controller and 24 for the other, increasing the RG on controller 2 to 23 (1 S, 2 P, 20 D)???

Just looking for some advice and ideas?

[vmsjaak13]

In your case 1 spare per controller is fine.

How to divide your disks is ultimately up to you. I.e. 12/24 if you need more performance/capacity on the other controller.

If you ever need to upgrade the 2040 to a 32xx system, you can't reuse the internal disks, in which case 12/24 is easier to migrate.

You can increase the size of the raidgroup to 17 (max is 28).

Do this before adding disks, or else ontap will try to create a 2nd raidgroup, in which case you will end up with uneven RG's.

Regards,

Niek

[nsitps1976]

Will increasing the RG by a mere one disk have any negative impact?

Also, if you buy a 2040HA both root vols will be located within the unit - If at a later date you buy a 4324 can you move the root vol of one controller to the tray so you can have a 12/24 split?

[radek_kubka]

Hi,

Here is a very informative thread about RG sizes:

http://communities.netapp.com/message/2676#2676

Bigger RG size yields more usable capacity, but means longer rebuild times in a case of a disk failure.

This leads us back to the 1 vs. 2 hot-spares dilemma:

- 1 hot-spare means (obviously) more capacity

- 2 hot-spares are required to enable Maintenance Center:

"If Data ONTAP drive health monitoring determines that a drive has surpassed an error threshold, rapid RAID recovery is initiated to a hot spare. [...] Maintenance Center requires that at least two hot spares are present in the system. NetApp recommends using Maintenance Center to provide additional intelligence for drive error discovery and correction."

Regards,
Radek