Monitoring performance of a Vmware physical mode RDM Perfmon

We have an ESSbase server that is running within our Vmware virtual environment. Lately there have been some performance issues, the vendor seems to think that it could be tied to the storage on which the data lives.

From within windows on the guest OS running perfmon during slow performance we noticed that the counter for Disk Transfers/sec was averaging about 40 and sometimes as high as 200. From everything Ive seen online this points to a disk bottleneck.

This server has an RDM attached which is comprised of 4 15k sas drives in a RAID 10 configuration from our EMC VNX SAN so it should be performing better than this.

Im just looking for some advice on what I can monitor through vmware or windows that might provide some insight into the slowness.

Any help is appreciated.
LVL 35
Joseph DalyAsked:
Who is Participating?

Improve company productivity with a Business Account.Sign Up

x
 
Andrew Hancock (VMware vExpert / EE MVE^2)Connect With a Mentor VMware and Virtualization ConsultantCommented:
As this is a physical implementation, e.g. disk/drive is presented to the VM direct, you will need to look a VM statistics (perfmon).

But as an aside...

do you have a VMFS datastore, using more disks?

Checkout this:-

RDM versus VMDK performance

Conclusion: VMFS and RDM have similar performance. Don’t choose RDM for performance.
Source:http://www.vfrank.org/2011/03/22/performance-rdm-vs-vmfs/
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
If you have an RDM, it's passed through RAW to the VM.

So you will need to look at performance metrics within the VM.

4 x 15K SAS drives, in RAID 10, if you want more IOPS, you'll need to add more VMs.

Is it read or write, which is causing the issues?

How many IOPS do you require?
0
 
Joseph DalyAuthor Commented:
"4 x 15K SAS drives, in RAID 10, if you want more IOPS, you'll need to add more VMs."

 Im guessing you meant disks here. That is something we are currently looking into but would take time.

"Is it read or write, which is causing the issues?"
I waiting on the DBA to be available to run tests to determine.

So would you reccomend monitoring this using perfmon tools or the performance metrics in Vcenter.
0
Simplify Active Directory Administration

Administration of Active Directory does not have to be hard.  Too often what should be a simple task is made more difficult than it needs to be.The solution?  Hyena from SystemTools Software.  With ease-of-use as well as powerful importing and bulk updating capabilities.

 
Joseph DalyAuthor Commented:
I ran crystal diskmark against the drive in question that was having the poor performance. It looks like my 4k read and writes are terrible.

diskmark
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
What is the latency?

The way to increase this is too add more spindles to the RAID 10 set.

Is the database running slowly?

It's not a poor database or application, with no indexes, a poor designed database will always be a poor slow VM.

Can you move the VM to a local disks and test?
0
 
Joseph DalyAuthor Commented:
This is a Hyperion installation each update is taking approx A minute and a half to respond and be committed

Moving to a Nother DataStor on the VM might be a possibility but I won't be able to do that until I talk to the team responsible for managing this system

Adding more disks to that raid group is our last option since it will require a purchase of additional disk
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Not to fast, we've been testing a new product from PernixData.

http://www.pernixdata.com/

If you've never heard of them, they are going to take the Storage/VMware world over!

We no longer need to add any more disks to our SANs to increase latency and IOPS!

Simple idea, an SSD/PCI-e flash card in the host, (not in the SAN) so closer to the CPU reduces even more latency, read and writes are cached in the SSD, and trickled back to the storage.

ALL storage is supported. No need to add more disks to your LUN/SAN!

It's BRILLIANT, highly recommended!

and you can select which VMs benefit, and therefore size your cache accordingly.

It's better than ALL the other versions!

oh, and the two co-founders are two guys which left VMware, the guy which designed the VMFS storage system, and another member from the Storage Team.

So there know a little about VMFS Storage!

Before adding more disks, I would look at this solution above, we've always been told by SAN vendors, add more spindles, add more spindles to increase IOPS and Latency, or purchase SSDs, or purchase Flash Cache Head.....

now it's changed.....enter PernixData.

PS I do not work for PernixData, I've just used and tested the product!

It's Excellent!
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.