WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

[Xen-devel] RE: VCPU time reporting!

To: Xen developer <xen-devel@xxxxxxxxxxxxxxxxxxx>
Subject: [Xen-devel] RE: VCPU time reporting!
From: Ata Bohra <ata.husain@xxxxxxxxxxx>
Date: Fri, 5 Jun 2009 20:22:09 +0530
Delivery-date: Fri, 05 Jun 2009 07:52:35 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
Importance: Normal
In-reply-to: <C64DC23F.78B5%keir.fraser@xxxxxxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
References: <C64DC23F.78B5%keir.fraser@xxxxxxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Dear All,
 
I was trying to understand the increased CPU utilization percentage reported by xentop on ay domain (no matter domains waits at getty login). I was able to find that 'nash-hotlplug' process uses almost 100 % of the CPU if system is idle, but being a nice process if you kickoff some job, it relinguishes most of the cpu.
 
By adding 'pkill -9 nash-hotplug' in /etc/rc.local helped me get the normal values of CPU utilization. Just sharing in case some one else also lands in same situation as mine.
 
Thanks!
Ata
  

From: ata.husain@xxxxxxxxxxx
To: xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: VCPU time reporting!
Date: Thu, 4 Jun 2009 23:14:29 +0530

Dear All,
 
I am trying to extract the CPU time used by each domain. As per the xentop code, it reports the VCPU time using 'xenstat_vcpu_ns()'. I have a strange observation, I started a ubuntu-8 domain and left it without running any application at the getty login prompt, when I observe the xentop data, it still gives me a constant increase in the VCPU time for both the domain0 and domain1(ubuntu-8).
 
Just to further analyze the data, I ran a large matrix multiplication program on dom0 and expected that the increase in VCPU time for domain0 should increase substantially that domain1 (which is doing nothing), BUT it is observed that the increase in VCPU time is in ration os maximum 1:2 with respect to domain1: domain0.
 
Is there any reason for a idle domain to consume so much of CPU time, I have not set any cap for scheduling thus I observe the domain0 maximum utilization not more than 50% whereas domain1 always shows the CPU utililzation of >= 50%.
 
Is there a more cleaner way to get individual domain CPU utlization parameters.
 
Thanks for your support!
 
Regards,
Ata E Husain
University at Buffalo
 


Get easy photo sharing with Windows Live™ Photos. Drag n’ drop

Videos Get the latest video streams on movies, Try it!
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
<Prev in Thread] Current Thread [Next in Thread>