VMware Cloud Community
christianZ
Champion
Champion

Open unofficial storage performance thread

Attention!

Since this thread is getting longer and longer, not to mention the load times, Christian and I decided to close this thread and start a new one.

The new thread is available here:

Oliver Reeh[/i]

[VMware Communities User Moderator|http://communities.vmware.com/docs/DOC-2444][/i]

My idea is to create an open thread with uniform tests whereby the results will be all inofficial and w/o any

warranty.

If anybody shouldn't be agreed with some results then he can make own tests and presents

his/her results too.

I hope this way to classify the different systems and give a "neutral" performance comparison.

Additionally I will mention that the performance is one of many aspects to choose the right system.

The others could be e.g.

\- support quality

\- system management integration

\- distribution

\- self made experiences

\- additional features

\- costs for storage system and infrastructure, etc.

There are examples of IOMETER Tests:

=====================================

\######## TEST NAME: Max Throughput-100%Read

size,% of size,% reads,% random,delay,burst,align,reply

32768,100,100,0,0,1,0,0

\######## TEST NAME: RealLife-60%Rand-65%Read

size,% of size,% reads,% random,delay,burst,align,reply

8192,100,65,60,0,1,0,0

\######## TEST NAME: Max Throughput-50%Read

size,% of size,% reads,% random,delay,burst,align,reply

32768,100,50,0,0,1,0,0

\######## TEST NAME: Random-8k-70%Read

size,% of size,% reads,% random,delay,burst,align,reply

8192,100,70,100,0,1,0,0

The global options are:

=====================================

Worker

Worker 1

Worker type

DISK

Default target settings for worker

Number of outstanding IOs,test connection rate,transactions per connection

64,ENABLED,500

Disk maximum size,starting sector

8000000,0

Run time = 5 min

For testing the disk C is configured and the test file (8000000 sectors) will be created by

first running - you need free space on the disk.

The cache size has direct influence on results. By systems with cache over 2GB the test

file should be increased.

LINK TO IOMETER:

Significant results are: Av. Response time, Av. IOS/sek, Av. MB/s

To mention are: what server (vm or physical), Processor number/type; What storage system, How many disks

Here the config file *.icf

\####################################### BEGIN of *.icf

Version 2004.07.30

'TEST SETUP ====================================================================

'Test Description

IO-Test

'Run Time

' hours minutes seconds

0 5 0

'Ramp Up Time (s)

0

'Default Disk Workers to Spawn

NUMBER_OF_CPUS

'Default Network Workers to Spawn

0

'Record Results

ALL

'Worker Cycling

' start step step type

1 5 LINEAR

'Disk Cycling

' start step step type

1 1 LINEAR

'Queue Depth Cycling

' start end step step type

8 128 2 EXPONENTIAL

'Test Type

NORMAL

'END test setup

'RESULTS DISPLAY ===============================================================

'Update Frequency,Update Type

4,WHOLE_TEST

'Bar chart 1 statistic

Total I/Os per Second

'Bar chart 2 statistic

Total MBs per Second

'Bar chart 3 statistic

Average I/O Response Time (ms)

'Bar chart 4 statistic

Maximum I/O Response Time (ms)

'Bar chart 5 statistic

% CPU Utilization (total)

'Bar chart 6 statistic

Total Error Count

'END results display

'ACCESS SPECIFICATIONS =========================================================

'Access specification name,default assignment

Max Throughput-100%Read,ALL

'size,% of size,% reads,% random,delay,burst,align,reply

32768,100,100,0,0,1,0,0

'Access specification name,default assignment

RealLife-60%Rand-65%Read,ALL

'size,% of size,% reads,% random,delay,burst,align,reply

8192,100,65,60,0,1,0,0

'Access specification name,default assignment

Max Throughput-50%Read,ALL

'size,% of size,% reads,% random,delay,burst,align,reply

32768,100,50,0,0,1,0,0

'Access specification name,default assignment

Random-8k-70%Read,ALL

'size,% of size,% reads,% random,delay,burst,align,reply

8192,100,70,100,0,1,0,0

'END access specifications

'MANAGER LIST ==================================================================

'Manager ID, manager name

1,PB-W2K3-04

'Manager network address

193.27.20.145

'Worker

Worker 1

'Worker type

DISK

'Default target settings for worker

'Number of outstanding IOs,test connection rate,transactions per connection

64,ENABLED,500

'Disk maximum size,starting sector

8000000,0

'End default target settings for worker

'Assigned access specs

'End assigned access specs

'Target assignments

'Target

C:

'Target type

DISK

'End target

'End target assignments

'End worker

'End manager

'END manager list

Version 2004.07.30

\####################################### ENDE of *.icf

TABLE SAMPLE

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

TABLE oF RESULTS

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

SERVER TYPE: VM or PHYS.

CPU TYPE / NUMBER: VCPU / 1

HOST TYPE: Dell PE6850, 16GB RAM; 4x XEON 51xx, 2,66 GHz, DC

STORAGE TYPE / DISK NUMBER / RAID LEVEL: EQL PS3600 x 1 / 14+2 Disks / R50

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read........__________..........__________.........__________

RealLife-60%Rand-65%Read......__________..........__________.........__________

Max Throughput-50%Read..........__________..........__________.........__________

Random-8k-70%Read.................__________..........__________.........__________

EXCEPTIONS: CPU Util.-XX%;

##################################################################################

I hope YOU JOIN IN !

Regards

Christian

A Google Spreadsheet version is here:

Message was edited by:

ken.cline@hp.com to remove ALL CAPS from thread title

Message was edited by:

RDPetruska

Added link to Atamido's Google Spreadsheet

Tags (1)
0 Kudos
457 Replies
JUSAO
Contributor
Contributor

SERVER TYPE: VM ON ESX 3.5 / 2 GB / Windows 2003 ENT / VM WITH MS softmirror.

CPU TYPE / NUMBER: VCPU / 1

HOST TYPE: HP DL360R05, 14GB RAM; 2x XEON E5335 QC, 2,00 GHz,

STORAGE TYPE / DISK NUMBER / RAID LEVEL: 2 x HP AiO1200 SATA, 4Gt/ HP P800 / 11+1 spare / RAID 5

SAN TYPE / HBAs : iSCSI; MS & VMWARE iSCSI softw. initiator, 2XGb NIC for iSCSI

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read........___134____..........___439____.........___13,74__

RealLife-60%Rand-65%Read......___51,6___..........___433____.........____3,39__

Max Throughput-50%Read..........___193____..........___196,6__.........____6,14__

Random-8k-70%Read.................___51,02___..........___427,6__.........____3,34__

EXCEPTIONS: CPU Util.85-90%, AiO Cpu util: 5-10%, disk util: 2-5%, mem util: 2-5%

##################################################################################

SERVER TYPE: VM ON ESX 3.5 / 2 GB / Windows 2003 ENT / VM WITHOUT MS softmirror.

CPU TYPE / NUMBER: VCPU / 1

HOST TYPE: HP DL360R05, 14GB RAM; 2x XEON E5335 QC, 2,00 GHz,

STORAGE TYPE / DISK NUMBER / RAID LEVEL: HP AiO1200 SATA, 4Gt/ HP P800/ 11+1 spare / RAID 5

SAN TYPE / HBAs : iSCSI; MS & VMWARE iSCSI softw. initiator, 2XGb NIC for iSCSI

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read........___17,74____..........___3316,58_.........___103,64__

RealLife-60%Rand-65%Read......___2,63___..........___12582,67__.........___98,03__

Max Throughput-50%Read..........___11,84___..........___4983,24__.........__155,73__

Random-8k-70%Read.................___2,86___..........___12283,84__.........____95,97__

EXCEPTIONS: CPU Util. 31-95%, AiO Cpu util: 55-65%, disk util: 85-100%, mem util: 2-5%

##################################################################################

Huge difference with ios and MB/sec!!! I am not sure what to think about this because I did get too many ios in second test. With MS soft mirror AiO usage is minimum and I would get real time mirroring to two independent disk systems.

0 Kudos
larstr
Champion
Champion

Huge difference with ios and MB/sec!!! I am not sure what to think about this because I did get too many ios in second test. With MS soft mirror AiO usage is minimum and I would get real time mirroring to two independent disk systems.

I'm sorry, but I don't really understand your results. How could a 100% read from a RAID5 be 4 times as slow as a mixed workload from the same subsystem?

I'm also not sure I understand your setup. Both setups involve a P800 raid5 consisting of 11+1 sata disks. How does the softmirror fit in here?

Lars

0 Kudos
JUSAO
Contributor
Contributor

On first test vm uses MS softmirror/raid1 with two virtual disks on two independent HP AiO1200 (1) and HP AiO1200 (2). On second test without softmirror/raid1 only to one AiO1200 (1). Both AiO's has HP P800 raid controller with 11 active/1 spare disk RAID5.

0 Kudos
christianZ
Champion
Champion

Thanks for joining in - but my propose would be to verify your 2nd test (e.g. with esxtop/d) - especially the tests with ios more than 10000 (with 10 sata disks).

I guess the timing was incorrect there (during high cpu utilisation).

0 Kudos
JUSAO
Contributor
Contributor

I will try next week. System is allready in production use but I think I will still get results. But I think that MB/s is ok. You are right. You cannot get that much ios. Softraid works ok with load also.

0 Kudos
mcdmagix
Contributor
Contributor

SERVER TYPE: 2003 VM on 3.0.1

CPU TYPE / NUMBER: VCPU / 1

HOST TYPE: IBM x3650, 8GB RAM; 2x Xeon, 3GHz

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Local Storage, 6x 146GB 15k SAS, RAID 10

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read........_____10.5_____.........._____5644_____........._____176_____ 72% CPU

RealLife-60%Rand-65%Read......_____40_____.........._____1147_____........._____14_____ 43% CPU

Max Throughput-50%Read.........._____11_____..........____2501______........._____78_____ 80% CPU

Random-8k-70%Read................._____39_____..........____1214______.........____9______ 44% CPU

##################################################################################

0 Kudos
mitchellm3
Enthusiast
Enthusiast

SERVER TYPE: Win2003 Std. VM ON ESX 3.0.2 update1

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: HP DL585 G2, 32GB RAM, 4x AMD 8222 SE, 3.0 GHz, DC

STORAGE TYPE / DISK NUMBER / RAID LEVEL: IBM SVC 2145-8G4 in front of

DS4800 (7)shelves 4Gbit/sec 146GB drives. (1) 450GB VMFS / striped

SAN TYPE / HBAs : 4G FC Qlogic

*TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek---- *

Max Throughput-100%Read........____4.79____.........._11257___.........__351____

RealLife-60%Rand-65%Read......____2.47____.........._8457___.........___66____

Max Throughput-50%Read..........___0.89____.........._12199___.........___381____

Random-8k-70%Read.................___0.43____.........._16359___.........___127____

EXCEPTIONS: CPU 45%, 46%, 51%, 50%. without second proc, cpu is 99%

0 Kudos
servpure
Contributor
Contributor

SERVER TYPE: VM ESX 3.5

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: SuperMicro 6025B-URV 2 x Xeon E5420 2.5 Ghz QC 8GB RAM

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Local Storage / 8 x Seagate ES.2 500 GB / RAID 10 on LSI 8308ELP w/BBU

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read........___3.9___..........___12236____.........__382.4__.....40.7% CPU

RealLife-60%Rand-65%Read......___49.1___..........___1075___.........___8.4___.....18.3% CPU

Max Throughput-50%Read..........___30.1___..........___1755___.........__54.8____.....24.3% CPU

Random-8k-70%Read.................___58.6___..........___918___.........__7.2____.....15.8% CPU

##################################################################################

0 Kudos
christianZ
Champion
Champion

Wow - new test are there.

Thanks to:

mcdmagix

mitchellm3 (I see the SVC results the first time here - very impressive, it is really Enterprise level)

servpure

for joining in.

..and guys don't forget to verify your results with e.g. esxtop / d (because of not correct time cycles in vm - especially by high vcpu utilization)

0 Kudos
christianZ
Champion
Champion

By your results I wonder about the large difference between 100%Read and 50%Read seq. tests.

Check my test on page 21 (on Infortrend system). I would expect more than 300 MB/sec by 50%Read and lower response time.

The cpu utilization seems to be normal IMHO. I wouldn't think you can drop it down. On ph. server with only scsi controller you will see similar behavior - other on servers with raid cotrollers - they have own cpu for ios operations and this way unload the main server cpu.

Anyway verify your numbers with esxtop / d.

Is your raid controller caching enabled (with "write back")??

0 Kudos
s_buerger
Contributor
Contributor

Physical

SERVER

TYPE: Physical Win2k3 R2 Standard with MS Initiator 2.04

CPU TYPE / NUMBER: CPU / 2

HOST TYPE: HP DL380G5 - 4GB - 1x Intel Xeon E5335 QC 2,0 GHz

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic PS3700X / 16x 400GB 10k 3,5"

SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 2x GB NIC (1 Intel PT 1000 Dual Port) mpio round robin

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read 8.3977 7068.58 220.89

RealLife-60%Rand-65%Read 12.2540 3404.68 26.60

Max Throughput-50%Read 13.5661 3689.95 115.31

Random-8k-70%Read 13.1474 3272.06 25.56

Exceptions: No Jumbo Frames

##################################################################################

SERVER

TYPE: Physical Win2k3 R2 Standard with MS Initiator 2.06

CPU TYPE / NUMBER: CPU / 2

HOST TYPE: HP DL380G5 - 4GB - 1x Intel Xeon E5335 QC 2,0 GHz

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic PS3700X / 16x 400GB 10k 3,5"

SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 2x GB NIC (1 Intel PT 1000 Dual Port) mpio round robin

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek---Av. MB/sek----

##################################################################################

Max Throughput-100%Read 8.3680 7103.42 221.98

RealLife-60%Rand-65%Read 12.4885 3388.46 26.47

Max Throughput-50%Read 10.1285 5395.62 168.61

Random-8k-70%Read 13.2423 3239.76 25.31

Exceptions: No Jumbo Frames

##################################################################################

Virtual VMDK

All tests with ESX 3.0.2

SERVER TYPE: Win2k3 VM (1,0GB RAM, 200GB vmdk) on ESX

3.0.2

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: HP DL380G5 - 20GB - 2x Xeon5345 2.33GHz

Quadcore

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic

PS3700X / 16x 400GB 10k 3,5" SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 2x GB NIC (1 Intel PT 1000 Dual Port)

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek-----Av.

MB/sek------

##################################################################################

Max Throughput-100%Read 1.2736 21859.83 683.12

VMWare shows 45MB read and 27.000

Read Requests on that vmhba while benchmark running

RealLife-60%Rand-65%Read 36.2758 1486.94 11.62

Max Throughput-50%Read 13.5574 4099.93 128.12

Random-8k-70%Read 31.7119 1739.23 13.59

EXCEPTIONS: CPU Util.-50-23-30-19%

Exceptions:

No Jumbo Frames

##################################################################################

Virtual iSCSI Initiator (2.06)

SERVER TYPE: Win2k3 VM (1,0GB RAM, 500GB LUN) on ESX

3.0.2

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: HP DL380G5 - 20GB - 2x Xeon5345 2.33GHz

Quadcore

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic

PS3700X / 16x 400GB 10k 3,5" SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 1x vNIC on 2x pNIC

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek-----Av.

MB/sek------

##################################################################################

Max Throughput-100%Read 26.9306 2138.78 66.84

RealLife-60%Rand-65%Read 17.6637 2635.91 20.59

Max Throughput-50%Read 17.7519 2986.19 93.32

Random-8k-70%Read 19.0502 2474.21 19.33

EXCEPTIONS: CPU Util.-40-41-55-41%

Exceptions: No Jumbo Frames

##################################################################################

SERVER TYPE: Win2k3 VM (1,0GB RAM, 500GB LUN) on ESX

3.0.2

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: HP DL380G5 - 20GB - 2x Xeon5345 2.33GHz

Quadcore

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic

PS3700X / 16x 400GB 10k 3,5" SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 2x vNIC mpio round robin on 2x pNIC

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek-----Av.

MB/sek------

##################################################################################

Max Throughput-100%Read 14.3924 3763.21 117.60

RealLife-60%Rand-65%Read 19.0035 2423.66 18.93

Max Throughput-50%Read 12.1459 4372.56 136.64

Random-8k-70%Read 19.1890 2443.89 19.09

EXCEPTIONS: CPU Util.-61-38-72-40%

Exceptions: No Jumbo Frames

##################################################################################

SERVER TYPE: Win2k3 VM (1,0GB RAM, 500GB LUN) on ESX

3.0.2

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: HP DL380G5 - 20GB - 2x Xeon5345 2.33GHz

Quadcore

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic

PS3700X / 16x 400GB 10k 3,5" SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 3x vNIC mpio round robin on 2x pNIC

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek-----Av.

MB/sek------

##################################################################################

Max Throughput-100%Read 16.3571 3066.85 95.84

RealLife-60%Rand-65%Read 17.6417 2502.40 19.55

Max Throughput-50%Read 20.5898 2154.20 67.32

Random-8k-70%Read 18.9348 2449.84 19.14

EXCEPTIONS: CPU Util.-57-42-72-41%

Exceptions: No Jumbo Frames

##################################################################################

SERVER TYPE: Win2k3 VM (1,0GB RAM, 500GB LUN) on ESX

3.0.2

CPU TYPE / NUMBER: VCPU / 2

HOST TYPE: HP DL380G5 - 20GB - 2x Xeon5345 2.33GHz

Quadcore

STORAGE TYPE / DISK NUMBER / RAID LEVEL: Equallogic

PS3700X / 16x 400GB 10k 3,5" SAS / Raid 50

SAN TYPE / HBAs : iSCSI; 2x vNIC mpio least queue depth on 2x pNIC

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sek-----Av.

MB/sek------

##################################################################################

Max Throughput-100%Read 12.5303 4337.17 135.54

RealLife-60%Rand-65%Read 15.9793 2653.82 20.73

EXCEPTIONS: CPU Util.-50-38%

Exceptions: No Jumbo Frames

##################################################################################

p.s. this rich text editor sucks...

0 Kudos
doubleH
Expert
Expert

ChristianZ et al,

I need a different set of eyes to help me on this. I originally participated in this thread in July 2007 (pg 13). At the time I posted my IOMeter results for my new VI3 environment. At that time I learned of performance issues with running Jumbo's and Flow Control at the same time on some Procurve switches even though in the Procurve release notes they say it can be done. I brought this up to our Procurve SE and he offered to help test this.

To perform the testing I have a new server (DL385 G2, Dual Core Opteron 2.6ghz, 20gb RAM, QLA4062C HBA) and SAN (EQL PS100E) and HP has lent me Procurve 3500 to try and reproduce the jumbo slow downs. I am running ESX 3.5.0 U1 on this test server. So far I cannot get anywhere near my IOMeter posting back in July 2007. Either something is very wrong or I am not reading the CSV output correctly. Could someone help me out by looking at my attached results and tell me what my Av. Resp. Time ms----Av. IOs/sek-----Av. MB/sek are. These results are with Flow Control enabled on the HBA and SAN switch ports with a VM running W2K3 SP2 and IOMeter. HBA's have Jumbo's turned off.

If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
0 Kudos
rock0n
VMware Employee
VMware Employee

Ich befinde mich bis zum 25. April 2008 nicht im Haus.

Ich werde Ihre Mail so bald wie möglich bearbeiten.

Mit freundlichem Gruß

Raiko Mesterheide

WORTMANN AG, Bredenhop 20, 32609 Hüllhorst, Deutschland, Vorstand: Siegbert Wortmann, Aufsichtratsvorsitzender: Thomas Knicker

Amtsgericht Bad Oeynhausen, HRB 7371, Steuernummer 331 5886 0042, USt.-Id.-Nr.: DE 125 753 836, WEEE-Reg.-Nr. DE 267 235 92

0 Kudos
doubleH
Expert
Expert

Sorry dude....don't speak german

If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
0 Kudos
christianZ
Champion
Champion

What test are you running there?

'size,% of size,% reads,% random,delay,burst,align,reply

512,100,100,0,0,1,0,0




There isn't one of the thread's test any way. Copy the iometer config file (on first page, her the link:

) and you will get the right tests then.

0 Kudos
doubleH
Expert
Expert

Man I feel stupid. I have the same ICF with the same name, but now that

I re-downloaded the ICF the test results are looking MUCH better.

Thanks!

NOTE - I have a QLA4062C with the BETA firmware for the Jumbo's issue so

I am going to give this a shot and see what performance looks like.

Thanks Again!

If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
0 Kudos
doubleH
Expert
Expert

I've added a second EqualLogic PS100E to my array group and reran my IOMeter tests. My initial tests are included on page 13 of this thread. You will notice that the "Max Throughput 100% Read is almost the same, but the "RealLife and Random test results were better. Thinking this is a result of the way EqualLogic arrays scale. I was surprised to see that the "Max Throughput 50% Read" actually decreased. My results are attached as the formatting of the results looked horrible. I will be adding a 3rd array next week and will post the results.

If you found this or any other post helpful please consider the use of the Helpfull/Correct buttons to award points
0 Kudos
hicksj
Virtuoso
Virtuoso

First of all... congrats on this thread still rolling. Awesome info!

We recently acquired a DS3300 (iSCSI), and figured now is as good a time as any to run some tests... comparing to our old FC SAN.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ TABLE OF RESULTS – Win2003 SP2 / ESX 3.5 / DS3300 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ SERVER TYPE: VM CPU TYPE / NUMBER: VCPU / 1 HOST TYPE: HP DL380G5, 32GB RAM; 2 x E5450, 2.5GHz, Quad Core STORAGE TYPE / DISK NUMBER / RAID LEVEL: IBM DS3300 / 3 x 15K RPM SAS / 5 SAN TYPE / HBA: GigE / 2 x QLA4060C (Firmware: 3.0.1.24) ################################################################################## TEST NAME--


Av. Resp. Time ms--Av. IOs/sec
Av. MB/sec
- ################################################################################## Max Throughput-100%Read.......___5.9____..........___1176___.........____36____ RealLife-60%Rand-65%Read......___41_____..........___485____.........___3.7____ Max Throughput-50%Read........___9.5____..........___991____.........____31____ Random-8k-70%Read.............___32.9___..........___485____.........___3.7____ EXCEPTIONS:

CPU Util.- Each test under 15% Load on SAN – Essentially none. Only two hosts, one VM. Data consistent between Iometer & esxtop

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ TABLE OF RESULTS – Win2003 SP2 / ESX 3.0.2 / ESS800 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ SERVER TYPE: VM CPU TYPE / NUMBER: VCPU / 1 HOST TYPE: HP DL380G5, 20GB RAM; 2 x E5160, 3.0GHz, Dual Core STORAGE TYPE / DISK NUMBER / RAID LEVEL: IBM ESS800 / 8 / 5 SAN TYPE / HBA: FC (MDS9124) / 2 x LPE1150 ################################################################################## TEST NAME--


Av. Resp. Time ms--Av. IOs/sec
Av. MB/sec
- ################################################################################## Max Throughput-100%Read.......___18.2___..........__3203___.........___100____ RealLife-60%Rand-65%Read......___36.7___..........___491___.........___3.8____ Max Throughput-50%Read........___4.9____..........___981___.........____31____ Random-8k-70%Read.............___44.2___..........___490___.........___3.8____ EXCEPTIONS:

CPU Util.- 37%, 13%, 13%, 14%

Load on SAN – Production, many other active

Data consistent between Iometer & esxtop

##################################################################################

Honestly, I'm surprised to see that the results are so close. And, while I'm not trying to win any contests here, I'm somewhat concerned that my numbers seem quite low compared to some other results posted above. However, I notice that most results published above show very high vCPU utilization during their tests, and its possible the throughput numbers are quite skewed due to that. My results seem quite consistent with what esxtop was reporting during the tests, and my CPU utilization (according to Iometer & the OS) was pretty low. I'm really curious why the descrepency between my VM's cpu utilization vs. others...

Items I need to follow up on:

- test physical servers against both SANs

- create volumes utilizing higher number of disks on the DS3300

- try ESX software initiator instead of the qla4060

Regards, J

0 Kudos
hicksj
Virtuoso
Virtuoso

Well, I do like consistency... Here are results of a physical server hitting the DS3300 using a software initiator:

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

TABLE OF RESULTS – Win2003 SP2 Physical / DS3300

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

SERVER TYPE: Pysical

CPU TYPE / NUMBER: pCPU / 1

HOST TYPE: HP DL360G4, 4GB RAM; 3.0GHz XEON (HyperThread enabled)

STORAGE TYPE / DISK NUMBER / RAID LEVEL: IBM DS3300 / 3 x 15K RPM SAS / 5

SAN TYPE / HBA: GigE (No Jumbo) / MS Software Initiator

##################################################################################

TEST NAME--


Av. Resp. Time ms--Av. IOs/sec--


Av.

MB/sec---

##################################################################################

Max Throughput-100%Read.......___6.4____..........___1486___.........____46____

RealLife-60%Rand-65%Read......___37_____..........___468____.........___3.7____

Max Throughput-50%Read........___17_____..........___532____.........____17____

Random-8k-70%Read.............___26_____..........___486____.........___3.8____

EXCEPTIONS:

CPU Util.- 11%, 10%, 15%, 7%

Load on SAN – Essentially none. Only this physical host active.

##################################################################################

Other than the 3rd test (50%R), these values are very close to what our ESX3.5 hosted VM was experiencing. I was wondering whether the quite old firmware we're running (which is the only way I got the qla4060's to load in ESX) may have had an impact on throughput. According to this, that doesn't appear to be the case.

Now that I think of it, need to add another item to my list of followups - impact as concurrent IO sessions are added. That should be fun.

0 Kudos
christianZ
Champion
Champion

Thanks for joining in.

0 Kudos