iometer-user — For Iometer community (open discussion, helping each other etc.)

Re: [Iometer-user] iscsi

[Joe <de...@ei...>]

Nope. The OS has the initiator and the iscsi devices just show up as disk targets.

On July 24, 2020 2:48:22 PM CDT, jay...@gm... wrote:
>Does IOMETER support iscsi testing (does it setup the iSCSI Initiator)?
>
>
>
>-- 
>This email has been checked for viruses by AVG.
>https://www.avg.com

[Iometer-user] iscsi

[<jay...@gm...>]

Does IOMETER support iscsi testing (does it setup the iSCSI Initiator)?



-- 
This email has been checked for viruses by AVG.
https://www.avg.com

[Iometer-user] Iometer run time issue

[Dipongkar C <c.d...@mo...>]

HI
I am trying to get IOPS number of my ssd drive using IOmeter. I have done
all the settings as described in the manual.

Few doubts regarding,  how much time do I need to run the Io meter to get
the exact results. If i run the  IO meter for more than 15 minutes the IOPs
performance degrades.

Is there any workaround to fix this or Is there any specified time limit
to run the IOmeter. I can see while defining the sectors fix this issue.
But a bit confused if it will scan the whole drive or only the specified
sectors.

Also which is the best way to get IOPS physical or logical drive. Both
cases numbers are different.

-- 
Mobiveil INC., CONFIDENTIALITY NOTICE: This e-mail message, including any 
attachments, is for the sole use of the intended recipient(s) and may 
contain proprietary confidential or privileged information or otherwise be 
protected by law. Any unauthorized review, use, disclosure or distribution 
is prohibited. If you are not the intended recipient, please notify the 
sender and destroy all copies and the original message.

[Iometer-user] Iometer run time issue

[Dipongkar C <c.d...@mo...>]

Hi

I am trying to get IOPS number of my ssd drive using IOmeter. I have done
all the settings as described in the manual.

Only doubt is for how much time do I need to run the Io meter to get the
exact results. If i run the   IO meter for more than 15 minutes the IOPs
performance degrades.

Is there any workaround to fix this or Is there any specified time limit
to run the IOmeter.

Regards
Dipongkar.C

-- 
Mobiveil INC., CONFIDENTIALITY NOTICE: This e-mail message, including any 
attachments, is for the sole use of the intended recipient(s) and may 
contain proprietary confidential or privileged information or otherwise be 
protected by law. Any unauthorized review, use, disclosure or distribution 
is prohibited. If you are not the intended recipient, please notify the 
sender and destroy all copies and the original message.

[Iometer-user] iometer io submission pattern

[Saumyajit D. <sau...@gm...>]

Hi folks,

I am curious as to what io engine is used to submit IOs by iometer. Does it
use a native Windows IO submission engine like windowsaio by fio? Or does
it have any sort of optimisation?
The question arises due to the fact that I am getting way lesser read iops
while using fio as compared to iometer.

Regards,
Saumyajit Dey
Member of Technical Staff
VMware India

Re: [Iometer-user] Linux server no longer booting after iometer test

[Joe <de...@ei...>]

You did 50% read, which means the other 50% were writes, which would overwrite your selected raw disk.  If you want to do writes on a disk/partition/volume with data then you should run on the mounted file system. I liked to create my iobw.tst file myself first, so I can make it large enough so I don't just see caching.

You can run on 100% reads on a raw disk with out overwriting anything.

And just being nit picky, it didn't affect the partition table, sda2 still existed. I think it overwrote the superblock so your physical volumes were gone along with a whole bunch of writes over your volume groups, logical volume, file system.

I hope this helps you understand what happened. And I hope my info isn't too out of date, I haven't been involved with iometer for quite a while.


Joe



On November 13, 2018 6:07:51 AM CST, D'Haese Willem <Wil...@di...nt> wrote:
>Hello,
>
>We did some tests with iometer on Linux servers today and noticed that
>after running a 10 minute test with 32 workers, the Linux server is no
>longer able to boot. We selected sda2 as target.
>
>[cid:image001.png@01D47B50.2A9672D0]
>
>[cid:image002.png@01D47B51.220CF930]
>
>Is there something that can be done to prevent partition table problems
>after iometer tests on Linux?
>
>[cid:image005.jpg@01D47B51.E17D26F0]
>
>[cid:image006.png@01D47B51.E17D26F0]
>
>After running iometer, pvdisplay displays nothing. The server no longer
>recognizes it's disks.
>
>Grtz
>
>Willem D'Haese

[Iometer-user] Linux server no longer booting after iometer test

[D'Haese W. <Wil...@di...>]

Hello,

We did some tests with iometer on Linux servers today and noticed that after running a 10 minute test with 32 workers, the Linux server is no longer able to boot. We selected sda2 as target.

[cid:image001.png@01D47B50.2A9672D0]

[cid:image002.png@01D47B51.220CF930]

Is there something that can be done to prevent partition table problems after iometer tests on Linux?

[cid:image005.jpg@01D47B51.E17D26F0]

[cid:image006.png@01D47B51.E17D26F0]

After running iometer, pvdisplay displays nothing. The server no longer recognizes it's disks.

Grtz

Willem D'Haese

[Iometer-user] Iometer Automation

[Vivek P. <a1...@gm...>]

Hi,

We need to launch Iometer from GUI through automation.
Issue is, we are unable to find a way to run tests by clicking the Button
using automation.
Please suggest alternative approach for this.

*Note: *From command line we are able to launch the tests, but in our case
we need to launch tests from GUI only through automation.

Thanks

[Iometer-user] Trying to run IO Meter without GUI

[Draconas <aar...@gm...>]

Hello and thank you in advance for reading,

I am trying to run IOMeter using only a console window. I already have a IO
Meter config file so I will not need to set up any tests. I have attempted
to use the IOMeter source but nothing I have tried so far has worked. I have
been trying to use Visual Studio 2017 but after many days all I have is
error messages galore. I am stuck and do not now what the next step is or
even what I should be asking to get the correct answer. Does anyone have
tips and tricks I need to try in regards to what I trying to do? Any
relevant info is good info. Thank you for your time and consideration.

Have a fantastic day,

Draconas



--
Sent from: http://iometer.996274.n3.nabble.com/User-f3.html

[Iometer-user] Query Regarding IO-Meter IO Operations

[Guru P. R. A. <Gur...@wd...>]

Hi,

We have some queries with respect to IOMeter IO Operation.
Please find more information below.

Overview:
-            To Capture the Protocol trace while performing write operation on SD/uSD Cards using IOMeter with predefined. icf configuration, also verifying the Transfer Data Size (Block Size) value of each transfers defined in the. icf file.

Resources used:

-            IOMeter 2006.
-            Solid Gear v1.0 GUI and Hardware. (Capture trace).
-            OS: Windows 10 Pro

Problem Description:
-            Performed write Operation with parameter: Blocksize (4K, 16K, 32K, 64K etc), 100% Write, 100% Sequential, Queue Depth – 32 and Sector Size of 2097152 (1GB) using IOMeter on SD/uSD Card connected through Protocol Analyzer.
-            Sometimes Observed no write command is issued while performing write operation as noticed in solid gear and unable to capture the write trace for the same.
-            Sometimes when the write Command is observed in the SD protocol trace, the write command transfer chunk size is not matching the one specified in the Block size configuration.

Questions:

  *   How the Write operation is performed in IOMeter, does it transfer the entire 2097152 (1GB) of data to connected drive or does it creates small data on the drive and keeps on multiplying it till the size specified.
  *   Any suggestions regarding the tool to be used and the methodologies for testing such transfers and capturing the traces for IOMeter.


Thanks and Regards,
Guru Prasada Rao Amberkar
Principal Engineer, Test Engineering

Western Digital®
SanDisk India Device Design Centre Pvt. Ltd.
Prestige Tech Park, Prestige Excelsior,
Survey No.143/1 (Land), Amani Bellandur Khane Village,
Varthur Hobli, Bangalore East Taluk,
Bengaluru – 560 103.
Email:  gur...@wd...<mailto:gur...@wd...>
Office:  +91-80-4272-1433
Mobile: +91-9449204197

Re: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le

[<wei...@po...>]

Hi jose
    We use IBM Power8 CPU.
1. Makefile_Linux.ppc can support power8 planform server?
2. IOmeter-1.1.0 can complie on ubuntu16.04-server-ppc64le successfully?
3. I follow this bug and undefine max and min and can complie successfully，but when I dynamo to win64 and run program will show error info.
https://sourceforge.net/p/iometer/bugs/84/ 
Note: I try this under iometer-2006_07_27.common-src.(I use iometer-1.1.0-src again, 1.1.0 can not run, so do not show this error info.)



Weihua Zhu (朱卫华)
System Department
E-mail: wei...@po...
Tel: +86-512-68316328
Phone:13812765084
 
From: Jose F. Medeiros
Date: 2017-03-10 07:24
To: wei...@po...; iometer-user
Subject: Re: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le
I am not sure which dependency is causing your error, " In file included from Pulsar.cpp:135:0 " 
I don't think this guy got an answer either.
https://sourceforge.net/p/iometer/mailman/iometer-user/thread/OFA...@ag.../


Let me ask a dumb question, why are you using a PowerPC is it a Power Mac G5 ?


 
Sincerely, Jose F. Medeiros ,  408-256-0649 Google Voice 
San Jose, California
http://www.linkedin.com/in/josemedeiros
I have not failed. I've just found 10,000 ways that won't work. - Thomas A. Edison






From: "wei...@po..." <wei...@po...>
To: iometer-user <iom...@li...> 
Sent: Thursday, March 9, 2017 1:45 AM
Subject: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le

Hi IOmeter

 I can not complie iometer-1.1.0-src on ubuntu16.04-server-ppc64le.
Configure:
OS type: ubutu16.04-server for ppc64le
gcc is already the newest version (4:5.3.1-1ubuntu1)

Step:
root@ubuntu:~/iometer-1.1.0/src# cp Makefile-Linux.ppc Makefile
root@ubuntu:~/iometer-1.1.0/src# make all
/usr/bin/g++ -O3 -Wall -D_GNU_SOURCE -DIOMTR_OSFAMILY_UNIX -DIOMTR_OS_LINUX -DIOMTR_CPU_PPC -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_REENTRANT -DUSE_NEW_DISCOVERY_MECHANISM -DFORCE_STRUCT_ALIGN -c Pulsar.cpp
In file included from Pulsar.cpp:135:0:
/usr/include/c++/5/bits/sstream.tcc: In member function ‘virtual std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::overflow(std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type)’:
IOCommon.h:1041:33: error: expected unqualified-id before ‘(’ token
    #define max(a,b)             (((a) > (b)) ? (a) : (b))
                                 ^
IOCommon.h:1044:33: error: expected unqualified-id before ‘(’ token
    #define min(a,b)             (((a) < (b)) ? (a) : (b))
                                 ^
Makefile:143: recipe for target 'Pulsar.o' failed
make: *** [Pulsar.o] Error 1

 Ask:
    do you have any suggestion for me?
Thanks



Weihua Zhu (朱卫华)
System Department
E-mail: wei...@po...
Tel: +86-512-68316328
Phone:13812765084
------------------------------------------------------------------------------
Announcing the Oxford Dictionaries API! The API offers world-renowned
dictionary content that is easy and intuitive to access. Sign up for an
account today to start using our lexical data to power your apps and
projects. Get started today and enter our developer competition.
http://sdm.link/oxford_______________________________________________
Iometer-user mailing list
Iom...@li...
https://lists.sourceforge.net/lists/listinfo/iometer-user

Re: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le

[<wei...@po...>]

Hi jose
    We use IBM Power8 CPU.
1. Makefile_Linux.ppc can support power8 planform server?
2. IOmeter-1.1.0 can complie on ubuntu16.04-server-ppc64le successfully?
3. I follow this bug and undefine max and min and can complie successfully，but when I dynamo to win64 and run program will show error info.
https://sourceforge.net/p/iometer/bugs/84/ 
Note: I try this under iometer-2006_07_27.common-src.(I use iometer-1.1.0-src again, 1.1.0 can not run, so do not show this error info.)


Weihua Zhu (朱卫华)
System Department
E-mail: wei...@po...
Tel: +86-512-68316328
Phone:13812765084
 
From: Jose F. Medeiros
Date: 2017-03-10 07:24
To: wei...@po...; iometer-user
Subject: Re: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le
I am not sure which dependency is causing your error, " In file included from Pulsar.cpp:135:0 " 
I don't think this guy got an answer either.
https://sourceforge.net/p/iometer/mailman/iometer-user/thread/OFA...@ag.../


Let me ask a dumb question, why are you using a PowerPC is it a Power Mac G5 ?


 
Sincerely, Jose F. Medeiros ,  408-256-0649 Google Voice 
San Jose, California
http://www.linkedin.com/in/josemedeiros
I have not failed. I've just found 10,000 ways that won't work. - Thomas A. Edison






From: "wei...@po..." <wei...@po...>
To: iometer-user <iom...@li...> 
Sent: Thursday, March 9, 2017 1:45 AM
Subject: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le

Hi IOmeter

 I can not complie iometer-1.1.0-src on ubuntu16.04-server-ppc64le.
Configure:
OS type: ubutu16.04-server for ppc64le
gcc is already the newest version (4:5.3.1-1ubuntu1)

Step:
root@ubuntu:~/iometer-1.1.0/src# cp Makefile-Linux.ppc Makefile
root@ubuntu:~/iometer-1.1.0/src# make all
/usr/bin/g++ -O3 -Wall -D_GNU_SOURCE -DIOMTR_OSFAMILY_UNIX -DIOMTR_OS_LINUX -DIOMTR_CPU_PPC -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_REENTRANT -DUSE_NEW_DISCOVERY_MECHANISM -DFORCE_STRUCT_ALIGN -c Pulsar.cpp
In file included from Pulsar.cpp:135:0:
/usr/include/c++/5/bits/sstream.tcc: In member function ‘virtual std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::overflow(std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type)’:
IOCommon.h:1041:33: error: expected unqualified-id before ‘(’ token
    #define max(a,b)             (((a) > (b)) ? (a) : (b))
                                 ^
IOCommon.h:1044:33: error: expected unqualified-id before ‘(’ token
    #define min(a,b)             (((a) < (b)) ? (a) : (b))
                                 ^
Makefile:143: recipe for target 'Pulsar.o' failed
make: *** [Pulsar.o] Error 1

 Ask:
    do you have any suggestion for me?
Thanks



Weihua Zhu (朱卫华)
System Department
E-mail: wei...@po...
Tel: +86-512-68316328
Phone:13812765084
------------------------------------------------------------------------------
Announcing the Oxford Dictionaries API! The API offers world-renowned
dictionary content that is easy and intuitive to access. Sign up for an
account today to start using our lexical data to power your apps and
projects. Get started today and enter our developer competition.
http://sdm.link/oxford_______________________________________________
Iometer-user mailing list
Iom...@li...
https://lists.sourceforge.net/lists/listinfo/iometer-user

Re: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le

[Jose F. M. <jos...@ya...>]

I am not sure which dependency is causing your error, " In file included from Pulsar.cpp:135:0 " 
I don't think this guy got an answer either.
https://sourceforge.net/p/iometer/mailman/iometer-user/thread/OFA...@ag.../

Let me ask a dumb question, why are you using a PowerPC is it a Power Mac G5 ?


 Sincerely, Jose F. Medeiros ,  408-256-0649 Google Voice 
San Jose, California
 http://www.linkedin.com/in/josemedeiros
 I have not failed. I've just found 10,000 ways that won't work. - Thomas A. Edison



      From: "wei...@po..." <wei...@po...>
 To: iometer-user <iom...@li...> 
 Sent: Thursday, March 9, 2017 1:45 AM
 Subject: [Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le
   
#yiv2087541039 body {line-height:1.5;}#yiv2087541039 body {font-size:10.5pt;color:rgb(0, 0, 0);line-height:1.5;}Hi IOmeter
 I can not complie iometer-1.1.0-src on ubuntu16.04-server-ppc64le.Configure:OS type: ubutu16.04-server for ppc64legcc is already the newest version (4:5.3.1-1ubuntu1)
Step:root@ubuntu:~/iometer-1.1.0/src# cp Makefile-Linux.ppc Makefileroot@ubuntu:~/iometer-1.1.0/src# make all/usr/bin/g++ -O3 -Wall -D_GNU_SOURCE -DIOMTR_OSFAMILY_UNIX -DIOMTR_OS_LINUX -DIOMTR_CPU_PPC -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_REENTRANT -DUSE_NEW_DISCOVERY_MECHANISM -DFORCE_STRUCT_ALIGN -c Pulsar.cpp
In file included from Pulsar.cpp:135:0:
/usr/include/c++/5/bits/sstream.tcc: In member function ‘virtual std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::overflow(std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type)’:
IOCommon.h:1041:33: error: expected unqualified-id before ‘(’ token
    #define max(a,b)             (((a) > (b)) ? (a) : (b))
                                 ^
IOCommon.h:1044:33: error: expected unqualified-id before ‘(’ token
    #define min(a,b)             (((a) < (b)) ? (a) : (b))
                                 ^
Makefile:143: recipe for target 'Pulsar.o' failed
make: *** [Pulsar.o] Error 1
 Ask:
    do you have any suggestion for me?Thanks
Weihua Zhu (朱卫华)System DepartmentE-mail: wei...@po...Tel: +86-512-68316328Phone:13812765084------------------------------------------------------------------------------
Announcing the Oxford Dictionaries API! The API offers world-renowned
dictionary content that is easy and intuitive to access. Sign up for an
account today to start using our lexical data to power your apps and
projects. Get started today and enter our developer competition.
http://sdm.link/oxford_______________________________________________
Iometer-user mailing list
Iom...@li...
https://lists.sourceforge.net/lists/listinfo/iometer-user


   

[Iometer-user] IOmeter 1.1.0 for ubuntu-16.04-server-ppc64le

[<wei...@po...>]

Hi IOmeter

 I can not complie iometer-1.1.0-src on ubuntu16.04-server-ppc64le.
Configure:
OS type: ubutu16.04-server for ppc64le
gcc is already the newest version (4:5.3.1-1ubuntu1)

Step:
root@ubuntu:~/iometer-1.1.0/src# cp Makefile-Linux.ppc Makefile
root@ubuntu:~/iometer-1.1.0/src# make all
/usr/bin/g++ -O3 -Wall -D_GNU_SOURCE -DIOMTR_OSFAMILY_UNIX -DIOMTR_OS_LINUX -DIOMTR_CPU_PPC -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_REENTRANT -DUSE_NEW_DISCOVERY_MECHANISM -DFORCE_STRUCT_ALIGN -c Pulsar.cpp
In file included from Pulsar.cpp:135:0:
/usr/include/c++/5/bits/sstream.tcc: In member function ‘virtual std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::overflow(std::__cxx11::basic_stringbuf<_CharT, _Traits, _Alloc>::int_type)’:
IOCommon.h:1041:33: error: expected unqualified-id before ‘(’ token
    #define max(a,b)             (((a) > (b)) ? (a) : (b))
                                 ^
IOCommon.h:1044:33: error: expected unqualified-id before ‘(’ token
    #define min(a,b)             (((a) < (b)) ? (a) : (b))
                                 ^
Makefile:143: recipe for target 'Pulsar.o' failed
make: *** [Pulsar.o] Error 1

 Ask:
    do you have any suggestion for me?
Thanks



Weihua Zhu (朱卫华)
System Department
E-mail: wei...@po...
Tel: +86-512-68316328
Phone:13812765084

[Iometer-user] why un-partitioned disk start at once

[Ming Ye (samye) <sa...@ci...>]

Hi all:

     I’m a user of iometer, and have a question on "disk target”

     I know that iometer will writing the data into the target disk then begin testing. But why the testing start so fast if I choose the unpartitioned disk(the blue icon)  and it seems that iometer does not write the data onto this disk.

    what’s the mechanism on writing the data to the unpartitioned disk? Tks a lot



Samye
Best regards

[Iometer-user] Headless Iometer

[Gordon B. <gor...@or...>]

Hello,

 

   I've tried searching via Google and various mailing lists but I have not been able to find an answer to my question "Is there a Headless version of Iometer available ?".  By headless I mean one which does not require a Windows GUI interface to be displayed.

 

    This may seem like a strange request but the way I am using Iometer is not to measure I/O performance but to generate an I/O workload while I perform various tests on a storage subsystem.  The way I have been running Iometer, until recently, has been to copy the Iometer and Dynamo executables to a Windows Server machine and then to execute Iometer supplying both a configuration and a results file on the command line so that there is no need for user interaction on the Windows system.  When the Iometer test run is completed the results file is retrieved from the Windows Server machine and analyzed on a Linux system.

 

    I initially had a lot of problems getting this to work because the Iometer GUI insisted on accessing a screen to display it's GUI interface.  I eventually managed to resolve this issue by using the Windows Sysinternals PsExec utility that allowed be to direct Iometer to use the primary console screen on the Windows Server for its GUI.  However, after migrating some of the Windows Servers I use to Windows Server 2012 R2 I can no longer get this to work, primarily because of security changes implemented in Windows Server 2012.

 

    It occurred to me that, in my situation, there was no real need for the Iometer GUI at all as it is never seen by the user and all the configuration and result information is maintained in files so I was wondering if anyone else has ever had a similar situation and created a headless version of Iometer that does not attempt to interact with a Windows screen at all.

 

    If anyone knows of such a version of Iometer, or a way to achieve the same effect on a Windows Server 2012 R2 system, that I was previously doing with PsExec, I would be very interested in hearing the details.  Many thanks,

 

      Gordon

 

[Iometer-user] Error 21 writing 4096 bytes

[Ankit M. <ank...@gm...>]

Hello,

I am using IOmeter to test whether MPIO on windows 2012 works  properly or
not, for that I run IOmeter trigger path failover (which takes a a min or
so).

But Iometer start to report below error.

 If I stop and start Iometer no issue is seen,

 but my expectation was to test MPIO failover with IOmeter running, IOmeter
should keep reporting IOs without any error count,

 but this is not happening. Can you please help what could be the problem

Thanks
Ankit

Error 21 writing 4096 bytes to disk
\\?\Volume{9af4a12d-d2fe-4d68-8559-e0a95dfd4491}.

[Iometer-user] replay test

[Ata F. b. <afb...@gm...>]

hi all
i found no forum to ask my question. so i decided to ask here.
sorry if my question is low level!
im new to IOmeter.
with a trace i wanna to

(1) increase data reuse level and
(2)  replay its past 300 seconds’ access traces

how can i do it with IOmeter ?!
thank a lot.

Re: [Iometer-user] "iometer alignment value is not a multiple of the sector size"

[Karim G. <kg...@2c...>]

hi all

Sorry for my late answer but i had to finish some important NVMe performance tests.


For now with 10 NVMe P3700 with 512GB memory (96*16) on a quad cpu motherboard with 4 E7-8890v3 I got:


32.45 GB/s en lecture séquentielle 4K

13.20 GB/s en écriture séquentielle 4K

3.2 Million IOPS en lecture Random

1.1 Million IOPS en écriture Random


Each cpu has 18Cores/36Threads using Windows server 2012 R2 datacenter edition i get some issues about the number of detected cpu by the OS and by IOMETER.


when i check #cpu option in iometer it display only 64 disk workers.

in the device manager all the 144 logical cores are listed

i tried to set 144 workers to have 1 worker per logical core or at least 72 workers like the number of physical cores .


i heard that its possible to go over what i got like results.

that's why i need help to set perfectly iometer and my HW to get the best results.


Can someone give me some explanations or best practice ?

how many outstanding io should i set when i use 72 workers or should i change the number of disk workers ?


Like you see im not sure at all.


warm regards






Karim GASMI
R&D - Support technique
www.2crsi.com<http://www.2crsi.com/>
kg...@2c...<mailto:kg...@2c...>
skype:k-rim_g

32, Rue Jacobi -Netter
67200 Strasbourg
Tél : +33 (0)3 68 41 10 65
GSM +33(0)6 23 06 67 44
Fax : +33 (0)9 59 25 10 75


     Pensez à l'environnement. N'imprimez ce courriel que si vous en avez vraiment besoin.




________________________________
De : FRANK Michael <mic...@fa...>
Envoyé : lundi 19 octobre 2015 11:27
À : 'iom...@li...'
Objet : [Iometer-user] "iometer alignment value is not a multiple of the sector size"


Hello,



First I want to say hello to the list and many thanks to all the development done from the community in order to provide such a great tool like Iometer for free.



I currently run some tests with Iometer 1.1.0 on Linux against a NAS system with NFSv3 and NFSv4. I already ran similar tests under W2k8 with CIFS without problems



In my test setup I have 64kB Transfer Request Size and Align I/Os to 4kB.



As soon as I start the test I get a warning with following content:



"iometer alignment value is not a multiple of the sector size......"



I found a post at the mailing list from 2006 with an answer that there was already a similar post giving the answer but unfortunately I don't find the mentioned post.



Any help is much appreciated and even in case I could silently ignore this warning I would be happy to get som feedback in order to be sure that it is not impacting the test results.



Many thanks in advance



Michael

This electronic transmission (and any attachments thereto) is intended solely for the use of the addressee(s). It may contain confidential or legally privileged information. If you are not the intended recipient of this message, you must delete it immediately and notify the sender. Any unauthorized use or disclosure of this message is strictly prohibited.  Faurecia does not guarantee the integrity of this transmission and shall therefore never be liable if the message is altered or falsified nor for any virus, interception or damage to your system.

[Iometer-user] "iometer alignment value is not a multiple of the sector size"

[FRANK M. <mic...@fa...>]

Hello,

First I want to say hello to the list and many thanks to all the development done from the community in order to provide such a great tool like Iometer for free.

I currently run some tests with Iometer 1.1.0 on Linux against a NAS system with NFSv3 and NFSv4. I already ran similar tests under W2k8 with CIFS without problems

In my test setup I have 64kB Transfer Request Size and Align I/Os to 4kB.

As soon as I start the test I get a warning with following content:

"iometer alignment value is not a multiple of the sector size......"

I found a post at the mailing list from 2006 with an answer that there was already a similar post giving the answer but unfortunately I don't find the mentioned post.

Any help is much appreciated and even in case I could silently ignore this warning I would be happy to get som feedback in order to be sure that it is not impacting the test results.

Many thanks in advance

Michael
This electronic transmission (and any attachments thereto) is intended solely for the use of the addressee(s). It may contain confidential or legally privileged information. If you are not the intended recipient of this message, you must delete it immediately and notify the sender. Any unauthorized use or disclosure of this message is strictly prohibited.  Faurecia does not guarantee the integrity of this transmission and shall therefore never be liable if the message is altered or falsified nor for any virus, interception or damage to your system.

[Iometer-user] I/O performance of RAID SSD

[Jose F. M. <jos...@ya...>]

can any one run a test on the IOP performance of SSD when using RAID 0 / 5 and 10?

I'd like to see a publised performance baseline. Sincerely, Jose F. Medeiros
408-256-0649 Google Voice
San Jose, California
http://www.linkedin.com/in/josemedeiros

I have not failed. I've just found 10,000 ways that won't work. - Thomas A. Edison

Re: [Iometer-user] iometer 1.1.0 on Windows Hyper-V Server 2012

[Kun K. <KKa...@us...>]

Hi Marty

I ran into a similar problem in the past, the best I could do is just run the Dynamo on the Windows Core server and manage it from another Windows server or PC with a GUI.
Based on my experience the IOmeter manager can be scripted using the batch option but it fails if it cannot find the GUI console.

Kun Katsumata


-----Original Message-----
From: Marty O'Brien [mailto:ma...@mi...] 
Sent: Friday, February 06, 2015 10:19 AM
To: iom...@li...
Subject: [Iometer-user] iometer 1.1.0 on Windows Hyper-V Server 2012

I have been trying to run iometer on several "Windows Hyper-V Server 2012" and "Windows Hyper-V Server 2012 R2" machines ("without the GUI", although it is really just stripped down), and, when running the .exe from the command line (either the 64bit or 32bit installed versions) it just returns to the command line with no error and no explanation. I don't see anything in Windows error logs, either. It is basically doing nothing that I can see.

I briefly looked through the code but it is quite a bit of programming and I thought it would be better to come here first and see if anyone can help. I've done quite a few hours worth of research on "the Google" and can't find anyone having similar issues, and even found a few articles that suggest using iometer on this type of setup.

The machines are relatively stock, save a few group policy settings and Zabbix monitoring agent.

Does anyone know how to troubleshoot this, or is it "simply won't run" on a "Core Edition" based Windows distribution?

My next option would probably be to try a linux live-cd, but that would be very inconvenient as I cannot afford the downtime on any of these servers.

Thanks, in advance, for any advice!!

Marty O'Brien
IT Director
Midwest Petroleum 




This e-mail may contain confidential and privileged material for the sole use of the intended recipient. Any review, use, distribution or disclosure by others is strictly prohibited. 
If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender by reply e-mail and delete all copies of this message.

------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Iometer-user mailing list
Iom...@li...
https://lists.sourceforge.net/lists/listinfo/iometer-user

[Iometer-user] iometer 1.1.0 on Windows Hyper-V Server 2012

[Marty O'B. <ma...@mi...>]

I have been trying to run iometer on several "Windows Hyper-V Server 2012" and "Windows Hyper-V Server 2012 R2" machines ("without the GUI", although it is really just stripped down), and, when running the .exe from the command line (either the 64bit or 32bit installed versions) it just returns to the command line with no error and no explanation. I don't see anything in Windows error logs, either. It is basically doing nothing that I can see.

I briefly looked through the code but it is quite a bit of programming and I thought it would be better to come here first and see if anyone can help. I've done quite a few hours worth of research on "the Google" and can't find anyone having similar issues, and even found a few articles that suggest using iometer on this type of setup.

The machines are relatively stock, save a few group policy settings and Zabbix monitoring agent.

Does anyone know how to troubleshoot this, or is it "simply won't run" on a "Core Edition" based Windows distribution?

My next option would probably be to try a linux live-cd, but that would be very inconvenient as I cannot afford the downtime on any of these servers.

Thanks, in advance, for any advice!!

Marty O'Brien
IT Director
Midwest Petroleum 




This e-mail may contain confidential and privileged material for the sole use of the intended recipient. Any review, use, distribution or disclosure by others is strictly prohibited. 
If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender by reply e-mail and delete all copies of this message.

Re: [Iometer-user] IOMeter is not a valid Win32 application

[Larvoire, Jean-F. <jea...@hp...>]

Hello Wojtek,

I just tried it on an old Windows 2003 Standard x64 I have left, and it too has the same problem with iometer 1.1.0 Win32 and Win64 versions released on 2014-03-15.

The “Devel 1.1.0-ga” versions released on 2012-05-18 do work though, as do the versions I built myself from source.

I suspect the final 1.1.0 version posted on SourceForge was built to target WINVER >= 0x0600 (Windows Vista and Server 2008). This is probably the default for the most recent versions of Visual C++, if WINVER is left undefined.

So the best thing for you to do is to rebuild iometer from source, possibly adding a “#define WINVER 0x0500” ahead of the sources. Else if you don’t have Visual C++ installed, look for the older versions I mention, that do work on Windows 2003.

Jean-François


From: Aqz [mailto:po...@aq...]
Sent: Tuesday, 13 January, 2015 08:23
To: iom...@li...
Subject: [Iometer-user] IOMeter is not a valid Win32 application

Every time i try to run IOMeter on my Windows 2003 machines i get "IOMeter.exe is not a valid Win32 application".

To be exact they're Windows 2003 Standard x64 with SP2.

What can I do to make it work?

--
Wojtek

[Iometer-user] IOMeter is not a valid Win32 application

[Aqz <po...@aq...>]

Every time i try to run IOMeter on my Windows 2003 machines i get
"IOMeter.exe is not a valid Win32 application".

To be exact they're Windows 2003 Standard x64 with SP2.

What can I do to make it work?

-- 
Wojtek