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

RE: [Xen-devel] strange CPU utilization, could related to credit schedule ?



Hi George:
 
       I've been looking into the credit schedule over agian and again.
       Well, I not smart enough to get fully understanding.
       Could you help to clarify below understanding?
 
       1.  From the algorithm, since domains credits is  direct proportion to its weight,
I think if there are two cpu-bound domains with same weight, no matter how many
vcpus they have, they will have the same CPU times accmulated, right?
       2. if 1 is true, what the different between domains with same weight but have
different VCPUS(say one has 4 vcpus, another has 8)?
       3. I am fully understand the problems of "credit 1 schedule "in your ppt of "Xenschedulerstatus"
 
(1)Client hypervisors and audio/video ï
    Audio VM: 5% CPU
ï 2x Kernel-build VMs: 97% cpu
ï 30-40 audio skips over 5 minutes
 
Do you mean "kernel-build VMs" has great impact on "Audio VM", and does priority CSCHED_PRI_TS_BOOST
solve this?

(2)Not fair to latency-sensitive workloads
ï Network scp: âFair shareâ 50%, usage 20-30%
(3)ï Load balancing 64 threads (4 x 8 x 2)
ï Unpredictable
ï Not scalable
ï Power management, Hyperthreads

Could you help to explan more ?
 
many many thanks, those confusions really makes me headache, I am a bit of silly.
   
 
 

From: tinnycloud@xxxxxxxxxxx
To: george.dunlap@xxxxxxxxxxxxx; xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: RE: [Xen-devel] strange CPU utilization, could related to credit schedule ?
Date: Thu, 13 Jan 2011 12:29:05 +0800

85% is from VM.
I forget to tell that, 8VMS each of them has 2 VCPUS, and the 9th VM, which is 2008
has 8VCPUs. We are still trying to reproduce the scence.
 
I have questiones on VM idle. How does Xen know VM is idle, or when VM is idle,
what is VCPU state in Xen, blocked or runable, and how is the CPU utiliazation
calcauted?
(I assume that the Idle VM finish physical CPU use before the time splice,
and its state come to blocked, then put it into *inactive* queue, right?
But will it is possible VM's VCPU come back to *active* queue when VM still
in idle, then we may have the phenomenon of VCPU shift between twe queues?)
 
Also, when VM's load comes up, will its priority be set BOOST, thus put
the head of *active* queue to be sheduled earlier?

 
 
> Date: Wed, 12 Jan 2011 16:41:07 +0000
> Subject: Re: [Xen-devel] strange CPU utilization, could related to credit schedule ?
> From: George.Dunlap@xxxxxxxxxxxxx
> To: tinnycloud@xxxxxxxxxxx
> CC: xen-devel@xxxxxxxxxxxxxxxxxxx
>
> Where is that 85% number coming from -- is this from within the VM, or
> from xentop?
>
> If it's Windows reporting from within the VM, one hypothesis is that
> it has to do with processing and running with virtual time. It may
> simply be a side effect of the VM only getting a small percentage of
> the cpu.
>
> If it's xentop, it's probably the vm reacting somehow to getting only
> a small percentage of the CPU. We saw something like this with early
> versions of Windows 2k3, but that problem was addressed in later
> service packs. At any rate, to find out what Windows is doing would
> require a bit more investigation. :-)
>
> -George
>
> On Wed, Jan 12, 2011 at 2:41 PM, tinnycloud <tinnycloud@xxxxxxxxxxx> wrote:
> >  Hi Geogre:
> >
> >          We have quite strange CPU usage behaivor in one of our DomU(2008
> > HVM)
> >          Totally, our host has 16 physical CPU, and 9 VMS.
> >
> >          Most of time, the all VMs works fine, the CPU usage are low and
> > resonable,
> > But at every high workload time(say 9:00-11:00AM, there are 8 VMs, each is a
> > web server,
> > cutomers accesses the page at this time), we login into the 9th VM which
> > is idle, find that
> > its CPU usage is at 85%, doesn't make any sense since we have no task
> > running, also the
> > usage distrbutes evenly across most of the processes.
> >
> >         I wonder if it relates to CPU schedule algorithm in Xen.
> >         After go
> > through http://lists.xensource.com/archives/html/xen-devel/2010-07/msg00414.html
> >         I can't figure out any assumptiones explains our situation.
> >         So what do u think of this?
> >
> >         Many thanks.
> >
> >
> >
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@xxxxxxxxxxxxxxxxxxx
> > http://lists.xensource.com/xen-devel
> >
> >
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.