iometer result and iostat result doesn't match

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

iometer result and iostat result doesn't match

qingwei wei
Hi,

I am currently running IOmeter (2006.07.27) on Linux machine (with SAN Storage). When running IOmeter, i also run iostat to capture the disk output every 60 seconds. What i found interesting is when i checked the throughput (MBps), IOMeter is reporting a throughput of ~265MBps while iostat result report around 500MBps. 

For IOmeter (1.1.0), the GUI result is around 250MBps and iostats result report 128MBps.

Anyone encounter such issue when they run IOmeter on Linux machine?

cw

------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: iometer result and iostat result doesn't match

Carl Zwanzig-2
(Sorry for the top post, OWA doesn't give a choice)

Yes, I've seen this and similar problems with reported speeds.There were also problems with dynamo crashing. In the end, I gave up looking for the root cause and dropped iometer/linux from our test scheme.

z!
________________________________________
From: qingwei wei [[hidden email]]
Sent: Wednesday, November 06, 2013 12:00 AM
To: [hidden email]
Subject: [Iometer-user] iometer result and iostat result doesn't match

Hi,

I am currently running IOmeter (2006.07.27) on Linux machine (with SAN Storage). When running IOmeter, i also run iostat to capture the disk output every 60 seconds. What i found interesting is when i checked the throughput (MBps), IOMeter is reporting a throughput of ~265MBps while iostat result report around 500MBps.

For IOmeter (1.1.0), the GUI result is around 250MBps and iostats result report 128MBps.

Anyone encounter such issue when they run IOmeter on Linux machine?

cw

------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: iometer result and iostat result doesn't match

qingwei wei
Hi Carl,

I presume iostat result should be more accurate than IOmeter?

Probably have to try fio for Linux machine.

cw


On Thu, Nov 7, 2013 at 12:09 AM, Carl Zwanzig <[hidden email]> wrote:
(Sorry for the top post, OWA doesn't give a choice)

Yes, I've seen this and similar problems with reported speeds.There were also problems with dynamo crashing. In the end, I gave up looking for the root cause and dropped iometer/linux from our test scheme.

z!
________________________________________
From: qingwei wei [[hidden email]]
Sent: Wednesday, November 06, 2013 12:00 AM
To: [hidden email]
Subject: [Iometer-user] iometer result and iostat result doesn't match

Hi,

I am currently running IOmeter (2006.07.27) on Linux machine (with SAN Storage). When running IOmeter, i also run iostat to capture the disk output every 60 seconds. What i found interesting is when i checked the throughput (MBps), IOMeter is reporting a throughput of ~265MBps while iostat result report around 500MBps.

For IOmeter (1.1.0), the GUI result is around 250MBps and iostats result report 128MBps.

Anyone encounter such issue when they run IOmeter on Linux machine?

cw


------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: iometer result and iostat result doesn't match

Wesley_Vaske

Not necessarily. I’ve had issues with iostat in the past doubling the throughput numbers. This was specifically when testing SAN storage. IOMeter would agree with the SAN utility while iostat showed doubled values.

 

Wes Vaske

Systems Engineer

Dell | Oracle Solutions

 

From: qingwei wei [mailto:[hidden email]]
Sent: Wednesday, November 06, 2013 7:03 PM
To: Carl Zwanzig
Cc: [hidden email]
Subject: Re: [Iometer-user] iometer result and iostat result doesn't match

 

Hi Carl,

 

I presume iostat result should be more accurate than IOmeter?

 

Probably have to try fio for Linux machine.

 

cw

 

On Thu, Nov 7, 2013 at 12:09 AM, Carl Zwanzig <[hidden email]> wrote:

(Sorry for the top post, OWA doesn't give a choice)

Yes, I've seen this and similar problems with reported speeds.There were also problems with dynamo crashing. In the end, I gave up looking for the root cause and dropped iometer/linux from our test scheme.

z!
________________________________________
From: qingwei wei [[hidden email]]
Sent: Wednesday, November 06, 2013 12:00 AM
To: [hidden email]
Subject: [Iometer-user] iometer result and iostat result doesn't match


Hi,

I am currently running IOmeter (2006.07.27) on Linux machine (with SAN Storage). When running IOmeter, i also run iostat to capture the disk output every 60 seconds. What i found interesting is when i checked the throughput (MBps), IOMeter is reporting a throughput of ~265MBps while iostat result report around 500MBps.

For IOmeter (1.1.0), the GUI result is around 250MBps and iostats result report 128MBps.

Anyone encounter such issue when they run IOmeter on Linux machine?

cw

 


------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: iometer result and iostat result doesn't match

qingwei wei
Hi Wes,

Thanks for the comment. It seems i need to find another too to capture io stats on Linux machine. Anyone has experience using dstat?

cw


On Thu, Nov 7, 2013 at 9:19 AM, <[hidden email]> wrote:

Not necessarily. I’ve had issues with iostat in the past doubling the throughput numbers. This was specifically when testing SAN storage. IOMeter would agree with the SAN utility while iostat showed doubled values.

 

Wes Vaske

Systems Engineer

Dell | Oracle Solutions

 

From: qingwei wei [mailto:[hidden email]]
Sent: Wednesday, November 06, 2013 7:03 PM
To: Carl Zwanzig
Cc: [hidden email]
Subject: Re: [Iometer-user] iometer result and iostat result doesn't match

 

Hi Carl,

 

I presume iostat result should be more accurate than IOmeter?

 

Probably have to try fio for Linux machine.

 

cw

 

On Thu, Nov 7, 2013 at 12:09 AM, Carl Zwanzig <[hidden email]> wrote:

(Sorry for the top post, OWA doesn't give a choice)

Yes, I've seen this and similar problems with reported speeds.There were also problems with dynamo crashing. In the end, I gave up looking for the root cause and dropped iometer/linux from our test scheme.

z!
________________________________________
From: qingwei wei [[hidden email]]
Sent: Wednesday, November 06, 2013 12:00 AM
To: [hidden email]
Subject: [Iometer-user] iometer result and iostat result doesn't match


Hi,

I am currently running IOmeter (2006.07.27) on Linux machine (with SAN Storage). When running IOmeter, i also run iostat to capture the disk output every 60 seconds. What i found interesting is when i checked the throughput (MBps), IOMeter is reporting a throughput of ~265MBps while iostat result report around 500MBps.

For IOmeter (1.1.0), the GUI result is around 250MBps and iostats result report 128MBps.

Anyone encounter such issue when they run IOmeter on Linux machine?

cw

 



------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: iometer result and iostat result doesn't match

Carl Zwanzig-2
In reply to this post by Wesley_Vaske

Hi,

 

OTOH and IME :), also with storage systems, linux’s iostat agreed with the array’s info, the network traffic counters, and other tools running the same load. iometer was way off.

 

At this point, I pretty much use fio all the time unless there’s a customer/client requirement for iometer.

 

z!

 

From: [hidden email] [mailto:[hidden email]]
Sent: Wednesday, November 06, 2013 5:19 PM
To: [hidden email]; Carl Zwanzig
Cc: [hidden email]
Subject: RE: [Iometer-user] iometer result and iostat result doesn't match

 

Not necessarily. I’ve had issues with iostat in the past doubling the throughput numbers. This was specifically when testing SAN storage. IOMeter would agree with the SAN utility while iostat showed doubled values.

 

Wes Vaske

Systems Engineer

Dell | Oracle Solutions


------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Iometer-user mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/iometer-user
Loading...