Poor native io performance on kvm guest in Rhel 6.1

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi all,

 

I am having following test setup where I run multiple guest os on a host
machine on kvm.

 

 

Host machine  operating sysyetem- Rhel 6.1

Virtualization -KVM

Guest operating system -Rhel 6.1

Number of guest os - 2,4 ,8 ,16

Storage - Lvm based storage for guest os(disk storage)

RAM of guest machine-1 gb

 

Im testing IO benchmark to evaluate Disk io performance of kvm guests
using benchmark bonnie and iozone on 2 gb file size

So these benchmark tools are run on guest machines.

 

I tested the two benchmark results with following guest configuration
file 

 

1. <disk type='block' device='disk'>

      <driver name='qemu' type='raw' cache='none' io='threads'/>

 

2. <disk type='block' device='disk'>

      <driver name='qemu' type='raw' cache='none' io='native'/>

 

 

 

Benchmark results of bonnie for and iozone is better for io=threads
option as compared to io=native option

However redhat 6.1 technical notes say than we should keep io=native
option in guest configuration file for disk storage,because it gives
better IO results.

But in my case for file size of 2 gb performance of iothread is better
for disk storage.

 

So is the bug that was reported in rhel 6.0 fo native IO is still not
solved on rhel 6.1.

Please find the link below for looking at the bug that was found in rhel
6.0

https://bugzilla.redhat.com/show_bug.cgi?id=611205

 

Please find below data for comparison of io=thread and io=native option

 

1. Bonnie results tested on 2 gb filer size .ram size 1 gb, io=threads
option in guest xml file on kvm ,number of virtual machines = 8

 

 

Bonnie sequential output

 

Per Char(K/Sec)

 

49857.17

 

Block (K/Sec)

 

200353.38

 

Rewrite (K/Sec)

 

12028.58

 

 

Bonnie sequential input

 

Per Char(K/Sec)

 

26358.50

 

Block (K/Sec)

 

39596.83

 

 

 

2.  Bonnie results tested on 2 gb filer size .ram size 1 gb, io=native
option in guest xml file on kvm ,number of virtual machines = 8

 

 

Bonnie sequential output

 

Per Char(K/Sec)

 

48020.83

Block (K/Sec)

 

68927.33

Rewrite (K/Sec)

 

5150.92

 

Bonnie sequential input

 

Per Char(K/Sec)

 

6387.88

Block (K/Sec)

 

23258.08

 

 

As we can see from the output that performance of  io=threads is better
than io=native on lvm based disk storage.

But red hat says io=native is better.

But in my test setup 8 virtaual machine are running , so can somebody
throw some light that why red hat says native perfrmnace is better,
ithink the bug has not been solved in rhel 6.1 .because io=thread
ioption shows better results..

 

 

 

Regards

Amit Tewari

 




DISCLAIMER: 
----------------------------------------------------------------------------------------------------------------------- 
The contents of this e-mail and any attachment(s) are confidential and
intended 
for the named recipient(s) only.  
It shall not attach any liability on the originator or NECHCL or its 
affiliates. Any views or opinions presented in  
this email are solely those of the author and may not necessarily reflect the 
opinions of NECHCL or its affiliates.  
Any form of reproduction, dissemination, copying, disclosure, modification, 
distribution and / or publication of  
this message without the prior written consent of the author of this e-mail is 
strictly prohibited. If you have  
received this email in error please delete it and notify the sender 
immediately. . 
-----------------------------------------------------------------------------------------------------------------------
-- 
redhat-list mailing list
unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list


[Index of Archives]     [CentOS]     [Kernel Development]     [PAM]     [Fedora Users]     [Red Hat Development]     [Big List of Linux Books]     [Linux Admin]     [Gimp]     [Asterisk PBX]     [Yosemite News]     [Red Hat Crash Utility]


  Powered by Linux