Configuration of Hitachi AMS 2100 SAN Storage to VMWare ESX 3.5 with best performance

[MehulParmar]

We have setup of AMS 2100 Hitachi SAN Stroage consist of SATA 2 HDD with VMWare ESX 3.5 and below is the configuration

1. 1 TB of LUN with lose LUN concept at the storage side. (1 TB LUN = 250 GB x 4 LUN)

2. Two controller with 2 port on each so total 4 port from 2 controller.

2. Visible to ESX via FC Brocade switch. each port support 4 GBPS data transfer.

The problem we are facing is

1. If we run 40-50 machines on ESX Server till that time, storage is performing well. But once we acceed the limit of 50 there is a drastic performance degrade for each VM image running on storage and if we will go and check the storage the write pending rate is more then expected.

Please can anyone let me know what could be the best practice to use Hitachi AMS 2100 storage with VMWare. It will be much appriciated.

Thanks.

[austar]

Am interested to see if you have upgraded to Vsphere and if you are still experiencing this issue.

We have just purchased a AMS2100 and are transfering VM's from NSC to AMS2100 storage on Vsphere.

Would like to see if any pitfalls may be coming on way.

thanks

[toha]

How many VMs you place on a single datastore?

[austar]

we have 400gig luns, no luse's , no more than 13 vm per lun at most, but that is based on VM usage, most have 4 - 5 per lun.

also we have sas drives.

[toha]

Last comment had good point, do not use LUSE LUNs with HDS storage and ESX. I personally saw lots of problems with our NSC55 when it had LUSE LUNs as ESX datastores, VMotions did fail most of the time etc, there is KB article about LUSE LUNs causing excessive SCSI reservation conflicts.

[jaygriffin]

How did your AMS 2100 perform on vSphere?

[Seb__Bauer]

Hi,

the easiest way to find out why the perfomance is degraded is to have a look in the performance monitor of the AMS2100 which is in Advanced Settings of the Storage Navigator Modular 2.

Then you can analyse where is the problem. High write pending rates are poision for good I/O rates. High usage of single disks or raid groups as well.

Why did you use LUSE LUNs? This is just an advantage when you pick a LUN out of different RAID groups.

Regards

Sebastian

[MehulParmar]

I would like to share my experience with Hitachi AMS 2100 modular storage.

This storage works good with LUSE Lun concept. We implemented LUSE Luns with ESX 4.0 and atleast getting good performance. At least better then tradition LUN concept.

Basically I found that, Hitachi AMS 2100 Storage itself gives so much problem while working with VMware. You can't utilize maximum with combination of Hitachi AMS 2100 & VMware ESX.

ESX Activities are quite throughput intensive and Hitachi AMS 2100 Modular Storage fails to co-operate in this condition. Specially Write pending rate goes upto 70%-75% in some conditions.

Even if you check Trans. Rate in performance-monitor -> back-end, it shows hardly upto 9 to 15 MB/S with Write Trans. Rate of from 5-8 MB/S, which is really not a good number.

We almost struggle for 6 months to get it worked.

[beckhamk]

We use the older AMS line and it runs fantastic! I do believe that HDS does say DO NOT use LUSE luns with vmware. So that is most likely problems #1. The second issue may possibly be your pushing too much i/o from the vms for the sata drives. I would look into reconifguring your luns to use normal luns from the ams.