Want to win a PS4? Go Premium and enter to win our High-Tech Treats giveaway. Enter to Win

x
?
Solved

Performance comparison between iSCSI-VMFS vs. NFS data store ?

Posted on 2016-11-24
9
Medium Priority
?
1,290 Views
Last Modified: 2016-11-30
I had this question after viewing MPIO and Link Aggregation (LACP) difference for iSCSI Network ?.

Hi All,

Further to my quetion previously in this forum, I just wanted to know of which one of these implementation performed better in terms of SQL Server & File Server workload:

QNAP TVS-471
"QNAP NFS -> NFS Datastore -> Windows Server VM"
"QNAP iSCSI -> VMFS Datastore -> Windows Server VM"

If anyone have tested or experience in the above two IP-Storage network technology, please let me know which one is better ?

Which tool or steps to be used:

https://labs.vmware.com/flings/i-o-analyzer or https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1006821

Thanks in advance.
0
Comment
[X]
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 5
  • 4
9 Comments
 
LVL 124

Accepted Solution

by:
Andrew Hancock (VMware vExpert / EE MVE^2) earned 2000 total points
ID: 41900991
a very difficult question to answer. TEST and Prove your environment as too which is better for your workload.

The problem with your Qnap is its not really an iSCSI SAN, e.g. compared with a Dell Equallogic with Service Processor and controllers.

It's a NAS device, with an ISCSI layer, which is common to lower end units and some higher end units, which claim to do everything, e.g. NFS, iSCSI and CIFS. So that's an additional layer which can hinder performance, and then you store the VM on shared storage, so you've got latency in your network, to affect performance, and then you use a VMDK, stored on the datastore, which also reduces performance....

and then what Qnap have you got, how many disk, Speed of disks, SATA, SAS, 7.2k, 10k, 15k, any cache, SSD, Flash, RAID type e.g. RAID 10, RAID 6, RAID 5...

all affects performance, so as you can see so many parameters...

As to which is better, it depends on many things, Workload (e.g. VMs), SAN, Network, e.g. 1GBe, or 10GBe, is it compatible with ESXi/VMware vSphere.

Try with and without Jumbo Frames...

As for tools, you could spend a month project determining benchmarks, but is that really going to help you with your workload.

and I think you've missed an option

1. Windows Server VM on NFS Export (NFS)
2. Windows Server VM on iSCSI Datastore (VMFS - QNAP)
3. Windows Server VM with iSCSI LUN on QNAP (using Software Init inside VM)
4. Direct Access to Windows Shares on CIFS (QNAP).

Which one is fastest - test....

As for applications, here are some more

When we are producing results for clients, we use the following applications, build spreadsheets, and average the results

1.IO Meter
http://www.iometer.org/

2. CrystalDiskMark
http://crystalmark.info/software/CrystalDiskMark/index-e.html

3. HD Tach
http://www.majorgeeks.com/files/details/hdtach.html

4. Intel NAS Toolkit
http://www.intel.com/products/server/storage/NAS_Perf_Toolkit.htm

5. VDBench
http://sourceforge.net/projects/vdbench

6. VMware Fling IO Blazer
http://labs.vmware.com/flings/ioblazer

7. Jet Stress
http://www.microsoft.com/en-us/download/details.aspx?id=36849

8. SQLIO
http://www.microsoft.com/en-gb/download/details.aspx?id=20163

9. HDPARM
http://linux.about.com/od/commands/l/blcmdl8_hdparm.htm

10. dd for (Linix, Unix and Windows)
http://en.wikipedia.org/wiki/Dd_%28Unix%29

11. ATTO
http://www.attotech.com/disk-benchmark/

12. Bonnie++
http://www.coker.com.au/bonnie++/

13. FIO
http://www.linux.com/learn/tutorials/442451-inspecting-disk-io-performance-with-fio

14. hIOmon
http://www.hyperio.com/productsAndServices.htm

15. DiskTester
http://diglloydtools.com/manual/disktester-iops.html

and here IO Meter is often the benchmark to use.

As for your question, as to which to use BOTH and MORE, and take an average....
2
 
LVL 8

Author Comment

by:Senior IT System Engineer
ID: 41905183
Andrew,

The QNAP TVS-471 NAS (https://www.qnap.com/en-us/product/model.php?II=158) that I'm using is VMware certified.

So if I configure the LUN:

2x 4 TB HDD - RAID 1
2x 1 TB HDD - RAID 0

Does the RAID-0 will perform faster but no redundancy ?
0
 
LVL 124
ID: 41905299
The fact is VMware certified, does not make it perform any better.

Therotically, the fastest configuration you will obtain in terms of RAW IOPS in the box could be using RAID 0, but other RAID configurations may yield more, e.g. more spindles = more disks = more IOPS = more performance
1
Get your Conversational Ransomware Defense e‑book

This e-book gives you an insight into the ransomware threat and reviews the fundamentals of top-notch ransomware preparedness and recovery. To help you protect yourself and your organization. The initial infection may be inevitable, so the best protection is to be fully prepared.

 
LVL 8

Author Comment

by:Senior IT System Engineer
ID: 41905318
Ah I see, so what about RAID-1 ?

More spindle than RAID-0.
0
 
LVL 124

Assisted Solution

by:Andrew Hancock (VMware vExpert / EE MVE^2)
Andrew Hancock (VMware vExpert / EE MVE^2) earned 2000 total points
ID: 41905469
It comes down to IOPS calculations, based on RAID...

Some very rough quick calculations

Total IOPS = (DriveIOPS * #Drives) / (ReadRatio + (RAIDWritePenalty * WriteRatio))

RAID 10 6 x 7,200 RPM SATA drive - Approx 50 IOPS each disk. assuming 33% Writes

Total IOPS = (50 IOPS x 6 disks)/(.67 + 2*.33))
           = 225 IOPS

RAID 5 6 x 600 SAS 15k Drives - Approx 200 IOPS each disk. assuming 33% writes

Total IOPS = (200 IOPS x 6 disks)/(0.67+4*.33))
           = 603 IOPS

Difference = 378 IOPS

What are these disks - SATA ? 7,200rpm?
1
 
LVL 8

Author Comment

by:Senior IT System Engineer
ID: 41905470
Yes it is 7200rpm SATA 3, 6 GBps.
0
 
LVL 124

Assisted Solution

by:Andrew Hancock (VMware vExpert / EE MVE^2)
Andrew Hancock (VMware vExpert / EE MVE^2) earned 2000 total points
ID: 41905504
So basically a single disk will give you - 40-100 IOPS per datastore (MAX).

RAID 1 - 50-133 IOPS per datastore (MAX)

based on IOPS per SATA disk 40-100 IOPS per disk.

This is the absolute theoretical maximum if you had VMs being accessed from the NAS, NOT across the network, because they will also reduce performance.....
1
 
LVL 8

Author Closing Comment

by:Senior IT System Engineer
ID: 41908004
Thanks Andrew !
0
 
LVL 124
ID: 41908015
no problems
0

Featured Post

Concerto's Cloud Advisory Services

Want to avoid the missteps to gaining all the benefits of the cloud? Learn more about the different assessment options from our Cloud Advisory team.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

The following article is comprised of the pearls we have garnered deploying virtualization solutions since Virtual Server 2005 and subsequent 2008 RTM+ Hyper-V in standalone and clustered environments.
Your data is at risk. Probably more today that at any other time in history. There are simply more people with more access to the Web with bad intentions.
This tutorial will walk an individual through the process of installing of Data Protection Manager on a server running Windows Server 2012 R2, including the prerequisites. Microsoft .Net 3.5 is required. To install this feature, go to Server Manager…
This video shows you how easy it is to boot from ISO images for virtual machines with the ISO images stored on a local datastore on the ESXi host.

609 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question