HPE Storage Users Group

Hello 3parug!

Today I met interesting issue. When I tried to create new CPG it gave me only r1 option. Other RAIDs free space are being calculated as zero.

It shows like this

std1 cli% showspace -p -devtype FC -ha mag -t r5
--Estimated(MB)---
RawFree UsableFree
0 0
std1 cli% showspace -p -devtype FC -ha mag -t r6
--Estimated(MB)---
RawFree UsableFree
0 0
std1 cli% showspace -p -devtype FC -ha mag -t r1
---Estimated(MB)---
RawFree UsableFree
13637632 6818816
std1 cli%

Does anybody know what is the matter? The goal is to create new r5 CGP with 5+1 set size , or at least change already created one.

_________________
*****************************************
Sorry for possible mistakes, English is not my mothertongue.

What's the physical config, model, trays, disks and are any of the disks currently full?

Hi ailean!

it's 8440, 2 cages with six FC 1800GB and ten SSD 1920GB drives on each.
typical FC drive is like this:



UPD: I forgot to note, all CPGs I mentioned are FC-only

_________________
*****************************************
Sorry for possible mistakes, English is not my mothertongue.

There is a slight limit I think on small amounts of disks.

Each node is master of half the disks in a tray and by default the layout of raid sets is at a node level.

So at the moment each Node sees 3 x FC disks and 5 x SSD disks. So CPG set sizes have to fit into 3 or 5.

There was an update in one of the 3.2.2 releases which added some more flexibility, allowing the set sizes to be across Node Pairs I think but I've only seen that on our SSD based 8xxx boxes so don't know if it applies to FC as well. After we had that update I was able to move our volumes from R1 to R5 (we only had 6 SSDs per tray), before the update I couldn't create a R5 CPG either.

Just had a look at the release notes (3.2.2MU2), I think the feature was "Express layout (for SSD only)" and as I suspected it does state SSD only.

Couldn't see any enhancements to that feature up to MU6 at least so I think you may be limited to R1 on FC unless you add more disks.

EDIT: On better news just checked the release notes for 3.3.1 and it states Express Layout enhanced to support all drives, not just SSD.

So r1 is available just because there are only two chunks per set. Thanks, you helped a lot.

I am afraid there will be no approve for 3PAR OS Upgrade, then additional disks is only solution.

Thanks again, Ailean!

_________________
*****************************************
Sorry for possible mistakes, English is not my mothertongue.

Minimum supported config is 6x SSDs, 12x FC or 12x NL per node pair. If your 8440 has 4 nodes, you require a minimum of 24x FC drives for a supported configuration with any FC drive.

_________________
The views and opinions expressed are my own and do not necessarily reflect those of my current or previous employers.

The Updated Rule is:
All node pairs must have minimum of either 6 Fast Class, 6 SSD, or 12 NL drives
https://hpe.kx.com.tw/file/repository/a ... yi1ipz.pdf
Table 3
6x FC with R1 or R5 (2+1) should work anytime, for R5 (5+1) with 6 Disk you need the Express Layout Feature.
This came with 3.2.2 EMU2 for SSD and 3.3.1 for all Drives.
(The CPG have to be created with the new IOS)
So it depends, if you have 2 or 4 Node 8440 and on your Firmware.
Both we did not know.
And also respect this:
https://support.hpe.com/hpsc/doc/public ... cale=en_US

Hi Cali

it's 4x-node with 3.2.2.612 (MU4)+P51,P56,P59,P74,P84,P92,P95

and it behaves exactly as you say - R1 or R5 (2+1) are only available.

_________________
*****************************************
Sorry for possible mistakes, English is not my mothertongue.