From: Peter Lieven <pl@dlh.net>
To: Vadim Rozenfeld <vrozenfe@redhat.com>
Cc: Gleb Natapov <gleb@redhat.com>, David Cure <kvm@cure.nom.fr>,
Avi Kivity <avi@redhat.com>,
kvm@vger.kernel.org
Subject: Re: performance trouble
Date: Mon, 26 Mar 2012 19:00:32 +0200 [thread overview]
Message-ID: <4F70A0B0.80309@dlh.net> (raw)
In-Reply-To: <201203221138.15413.vrozenfe@redhat.com>
On 22.03.2012 10:38, Vadim Rozenfeld wrote:
> On Thursday, March 22, 2012 10:52:42 AM Peter Lieven wrote:
>> On 22.03.2012 09:48, Vadim Rozenfeld wrote:
>>> On Thursday, March 22, 2012 09:53:45 AM Gleb Natapov wrote:
>>>> On Wed, Mar 21, 2012 at 06:31:02PM +0100, Peter Lieven wrote:
>>>>> On 21.03.2012 12:10, David Cure wrote:
>>>>>> hello,
>>>>>>
>>>>>> Le Tue, Mar 20, 2012 at 02:38:22PM +0200, Gleb Natapov ecrivait :
>>>>>>> Try to add<feature policy='disable' name='hypervisor'/> to cpu
>>>>>>> definition in XML and check command line.
>>>>>>>
>>>>>> ok I try this but I can't use<cpu model> to map the host cpu
>>>>>>
>>>>>> (my libvirt is 0.9.8) so I use :
>>>>>> <cpu match='exact'>
>>>>>>
>>>>>> <model>Opteron_G3</model>
>>>>>> <feature policy='disable' name='hypervisor'/>
>>>>>>
>>>>>> </cpu>
>>>>>>
>>>>>> (the physical server use Opteron CPU).
>>>>>>
>>>>>> The log is here :
>>>>>> http://www.roullier.net/Report/report-3.2-vhost-net-1vcpu-cpu.txt.gz
>>>>>>
>>>>>> And now with only 1 vcpu, the response time is 8.5s, great
>>>>>>
>>>>>> improvment. We keep this configuration for production : we check the
>>>>>> response time when some other users are connected.
>>>>> please keep in mind, that setting -hypervisor, disabling hpet and
>>>>> only one vcpu
>>>>> makes windows use tsc as clocksource. you have to make sure, that your
>>>>> vm is not switching between physical sockets on your system and that
>>>>> you have constant_tsc feature to have a stable tsc between the cores
>>>>> in the same socket. its also likely that the vm will crash when live
>>>>> migrated.
>>>> All true. I asked to try -hypervisor only to verify where we loose
>>>> performance. Since you get good result with it frequent access to PM
>>>> timer is probably the reason. I do not recommend using -hypervisor for
>>>> production!
>>>>
>>>>> @gleb: do you know whats the state of in-kernel hyper-v timers?
>>>> Vadim is working on it. I'll let him answer.
>>> It would be nice to have synthetic timers supported. But, at the moment,
>>> I'm only researching this feature.
>> So it will take months at least?
> I would say weeks.
Is there a way, we could contribute and help you with this?
Peter
next prev parent reply other threads:[~2012-03-26 17:00 UTC|newest]
Thread overview: 74+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-01-23 8:28 performance trouble David Cure
2012-01-23 10:04 ` David Cure
2012-01-30 15:36 ` David Cure
2012-01-30 16:20 ` Brian Jackson
2012-01-30 16:51 ` David Cure
2012-01-30 17:21 ` Avi Kivity
2012-01-31 17:15 ` David Cure
2012-01-31 17:21 ` Avi Kivity
2012-02-02 10:41 ` David Cure
2012-02-03 8:59 ` David Cure
2012-02-05 9:38 ` Avi Kivity
2012-02-10 10:09 ` David Cure
2012-02-14 13:32 ` Avi Kivity
2012-02-14 13:40 ` Gleb Natapov
2012-02-16 8:55 ` David Cure
2012-02-16 9:01 ` Gleb Natapov
2012-02-17 8:59 ` David Cure
2012-02-17 10:07 ` Gleb Natapov
2012-02-17 14:01 ` David Cure
2012-02-19 9:13 ` Gleb Natapov
2012-02-22 16:33 ` David Cure
2012-02-22 16:58 ` David Cure
2012-02-23 8:38 ` Gleb Natapov
2012-03-16 10:13 ` David Cure
2012-03-19 10:51 ` Gleb Natapov
2012-03-20 9:32 ` David Cure
2012-03-20 9:45 ` Gleb Natapov
2012-03-20 11:18 ` David Cure
2012-03-20 12:38 ` Gleb Natapov
2012-03-21 11:10 ` David Cure
2012-03-21 17:31 ` Peter Lieven
2012-03-22 7:53 ` Gleb Natapov
2012-03-22 7:57 ` Peter Lieven
2012-03-22 8:35 ` David Cure
2012-03-22 8:33 ` David Cure
2012-03-22 8:50 ` Peter Lieven
2012-03-22 8:48 ` Vadim Rozenfeld
2012-03-22 8:52 ` Peter Lieven
2012-03-22 9:38 ` Vadim Rozenfeld
2012-03-26 17:00 ` Peter Lieven [this message]
2012-03-26 17:46 ` Vadim Rozenfeld
2012-03-26 17:52 ` Gleb Natapov
2012-03-26 18:36 ` Vadim Rozenfeld
2012-03-26 18:54 ` Peter Lieven
2012-03-26 20:11 ` Vadim Rozenfeld
2012-03-27 8:56 ` Gleb Natapov
2012-03-27 9:23 ` Vadim Rozenfeld
2012-03-27 9:24 ` Gleb Natapov
2012-03-27 9:26 ` Peter Lieven
2012-03-27 10:00 ` Gleb Natapov
2012-03-27 12:20 ` Peter Lieven
2012-03-27 12:26 ` Gleb Natapov
2012-03-27 12:28 ` Peter Lieven
2012-03-27 12:29 ` Gleb Natapov
2012-03-27 12:30 ` Peter Lieven
2012-03-27 14:06 ` Peter Lieven
2012-03-27 15:44 ` Vadim Rozenfeld
2012-03-27 15:58 ` Peter Lieven
2012-03-27 16:12 ` Vadim Rozenfeld
2012-03-27 16:16 ` Peter Lieven
2012-03-27 17:06 ` Vadim Rozenfeld
2012-03-28 8:38 ` Peter Lieven
2012-03-27 10:40 ` Vadim Rozenfeld
2012-03-27 10:49 ` Peter Lieven
2012-03-27 11:43 ` Vadim Rozenfeld
2012-03-27 14:44 ` Peter Lieven
2012-03-27 15:37 ` Vadim Rozenfeld
2012-03-27 15:39 ` Peter Lieven
2012-03-27 15:55 ` Vadim Rozenfeld
2012-03-22 8:31 ` David Cure
2012-03-22 8:47 ` Peter Lieven
2012-02-14 13:48 ` Vadim Rozenfeld
2012-02-16 9:10 ` David Cure
2012-02-17 15:27 ` David Cure
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=4F70A0B0.80309@dlh.net \
--to=pl@dlh.net \
--cc=avi@redhat.com \
--cc=gleb@redhat.com \
--cc=kvm@cure.nom.fr \
--cc=kvm@vger.kernel.org \
--cc=vrozenfe@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;
as well as URLs for NNTP newsgroup(s).