public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Gregory Esnaud <Gregory.ESNAUD@exfo.com>
To: Paolo Bonzini <pbonzini@redhat.com>,
	"kvm@vger.kernel.org" <kvm@vger.kernel.org>
Subject: RE: [Newbie question] Why usage of CPU inside VM and outside VM is different
Date: Tue, 7 Jan 2020 17:32:55 +0000	[thread overview]
Message-ID: <fc4a7dfc9d344facaed8d34adcff3fc4@exfo.com> (raw)
In-Reply-To: <a8fb7ace-f52e-3a36-1c53-1db9468404e6@redhat.com>

Hi Paolo,

Thanks for your quick answer.

Are you using only 2 CPUs at 100%, or are you using 9 CPUs at a total of 200%? 

Inside the VM, a top command show that we are using 2 CPU @ 100%: https://framadrop.org/r/gNf3erJVf6#t2HLIpzPHGtxHzHeLrnGGjsRSZZpmv9HPVZt/LIIP8s=
From the hypervisor (ie, *outside* the vm), a top command show that the VM consuming 9 cpu: https://framadrop.org/r/-FmrBhMGX5#q/6aZ8Fq2Mnuzz3+6F/pa83U+jzPox1hd984tw7DkuA=

So, if I'm understanding you correctly, an parameter of kernel of our VM is idel=poll that we should change?


Thanks again,

-----Message d'origine-----
De : Paolo Bonzini <pbonzini@redhat.com> 
Envoyé : mardi 7 janvier 2020 17:59
À : Gregory Esnaud <Gregory.ESNAUD@exfo.com>; kvm@vger.kernel.org
Objet : Re: [Newbie question] Why usage of CPU inside VM and outside VM is different

On 07/01/20 16:57, Gregory Esnaud wrote:
> 
> From an hypervisor (via top command) point of view our VM is consuming 
> 9 CPU (900%). This was reported by our platform provider.
> From a VM point of view we are consuming only 2 CPU (200%), with a top 
> also.
> 
> Our provider claim us to explain why we are consuming so much CPU.
> But we cannot troubleshoot the infra as it's not our responsibility.
> From our point of view, everything is ok.

Are you using only 2 CPUs at 100%, or are you using 9 CPUs at a total of 200%?  If the former, it is possible that you are using something like idle=poll, so the machine _is_ idle but still consuming 100% host CPU.
This is something that you should know, however.

If the latter, the host is probably doing a little bit of busy waiting to improve your performance.  The provider can disable it using the
kvm.halt_poll_ns=0 module parameter and there is nothing that you can do about it.  But it is unlikely that it causes the utilization to jump so much.

Paolo


  reply	other threads:[~2020-01-07 17:33 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-07 15:57 [Newbie question] Why usage of CPU inside VM and outside VM is different Gregory Esnaud
2020-01-07 16:58 ` Paolo Bonzini
2020-01-07 17:32   ` Gregory Esnaud [this message]
2020-01-07 17:36     ` Paolo Bonzini
2020-01-07 17:37       ` Gregory Esnaud
2020-01-07 18:18         ` Gregory Esnaud

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=fc4a7dfc9d344facaed8d34adcff3fc4@exfo.com \
    --to=gregory.esnaud@exfo.com \
    --cc=kvm@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox