Weird performance issues MSA 2000G3

[M1kkel1984]

Hello.

I have a MSA2000G3 FC with 24 10K disks. Its splitted up in 4 vdisks on 6 disks each, running raid 10 16K chunk.

1 brocadce FC switch, and 1 brocade HBA in each ESX server. 4 ESX servers total. 

Running an iometer test (from open unofficial performance test - 8KB/ 35W/65R on a virtual server, i get about 280 iops. Nothing else hitting the vdisk. 

If i instead run it off a virual server on a HP DL380G4 with 6 10K 2,5" disks in raid 10, i get almost 2000 IOPS on the exact same test. 

How does this make sense? One is a SAN, the other a DAS. 

When i look in vmware performance/disk i reguarly see some high latency spikes on the luns (on the MSA 2000G3). Is this a controller cpu thing, or a misconfigured switch) or something else?

Running latest firmware btw. 

Maybe one of you guys with a nearly same conf could try to run this..

[John Kufrovich]

By those MSA performance numbers it looks like a caching issue.

Can you telnet to the array and check if you have any orphaned cache data.

>show unwritable

If the command returns unwritable cache, then you need to determine if it that volume is still present or faulted.  If its not there any longer, you can clear the cache >clear cach