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

Re: [libvirt] [PATCH 06/10] vcpubandwidth: introduce two new libvirt APIs



On 2011-7-6 16:56, Nikunj A. Dadhania wrote:
> On Wed, 06 Jul 2011 14:56:27 +0800, Gui Jianfeng <guijianfeng cn fujitsu com> wrote:
>> On 2011-7-5 16:41, Nikunj A. Dadhania wrote:
>>> On Tue, 05 Jul 2011 15:06:06 +0800, Wen Congyang <wency cn fujitsu com> wrote:
>>>> At 07/04/2011 07:19 PM, Nikunj A. Dadhania Write:
>>>>> On Thu, 30 Jun 2011 11:13:18 +0800, Wen Congyang <wency cn fujitsu com> wrote:
>>>>>> We want to control bandwidth for each vcpu, so we can not use the
>>>>>> API virDomainSetSchedulerParameters(). Introduce two new APIs to
>>>>>> change and query bandwidth for each vcpu.
>>>>>>
>>>>> Will we have different cpu bandwidth for different vcpus?
>>>>>
>>>>> Something like this:
>>>>>
>>>>> vcpu1: 1000000/250000
>>>>> vcpu2: 1000000/500000
>>>>> vcpu3: 1000000/300000
>>>>> vcpu4: 1000000/400000
>>>>>
>>>>> IMO, that is not required, we can have a top level bandwitdh for the VM
>>>>> and then redistribute it among vcpus equally, without user knowing about
>>>>> it.
>>>>>
>>>>> Something like this:
>>>>>
>>>>> VM1(4vcpu) has to be throttled at 1CPU bandwidth using SetSchedParamters.
>>>>>
>>>>> Internally libvirt splits it equally:
>>>>> vcpu1: 1000000/250000
>>>>> vcpu2: 1000000/250000
>>>>> vcpu3: 1000000/250000
>>>>> vcpu4: 1000000/250000
>>>>>
>>>>> So why introduce VCPU level apis?
>>>>
>>>> Adam Litke said IBM's performance team nead to control cpu bandwidth for each
>>>> vcpu.
>>> Right, but we do not export that as a User API, that was my suggestion.
>>> We can internally control each vcpu's bandwidth, i.e. divide equally.
>>
>> Hmm, I heard that some server could run CPUs at different speed.
>> May be this patch can simulate this behavior. 
> That happens on my laptop as well, depending on the machine load CPU
> frequency is changed but it is done transparently.

I means explicitly CPU speed configuring. ;)

> 
> I am not sure if we are trying to simulate that here.

So why not leave the flexible interface here, and let users make
the decision?

Thanks,
Gui

> 
> Regards
> Nikunj
> 
> 



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