Storage Boards Cleanup
To make it easier to find information about HPE Storage products and solutions, we are doing spring cleaning. This includes consolidation of some older boards, and a simpler structure that more accurately reflects how people use HPE Storage.
Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

MSA2324 vDISK Recommendations

MCurtis_1
Occasional Contributor

MSA2324 vDISK Recommendations

Hi

We have a new MSA2000 (2324FC G2)with 24 300Gb SAS disks.

I've read the "best practice" information from the HP site. I'm happy enough with the notion of (evenly) spreading the vdisks and physical disks between the two controller; and avoiding creating lots of ridiculously small vdisks (to avoid wasting space).

However, I still have some questions over the best (general) way to set up the beast.I say "general" because I don't need to wring out every last possible ounce of performance or space. Unfortunately, the "Best Practice" guide really left more questions than it answered.

The array is connected to redundant SAN fabric switches; will be accessed simultaneously by around 6 - 10 hosts (with varying workloads) and I would imagine that each vdisk would need to have multiple volumes mapped to these various hosts.

I appreciate it is difficult to give specific advice but in general terms:

a) Are there are sweet-spots when it comes to sizing vdisks e.g. an optimal number of disks in a RAID-5 set?
b) From a performance point of view would it be better to have two large vdisks (12 disks each), or four slightly smaller vdisks (6 disk each). (NB: For the sake of argument I'm ignoring the need for hot spares).

Many thanks for any help or advice.

Mark
1 REPLY
Prokopets
Respected Contributor

Re: MSA2324 vDISK Recommendations

a) As for me, it's not a good idea to use raid-5 with more than 10 disks. Just because of a probability of two disk failure. If you have a hot-spare disk the risk is quite lower.
b) it depends :). Look, your OS LVM can be faster than MSA controller when it manages two LUNs, or can be slower, i think, it's unpredictable.

I think, that it's better to give up a theoretical discussion and just make some tests with iometer - it will give you much more adequate answer.