You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@cloudstack.apache.org by Abdul Rasool <ra...@gmail.com> on 2015/01/01 11:24:38 UTC

Re: CPU socket miss match !!! Urgent

Thanks for your support Kevin

I have fine tuned cpu.overprovisioning.factor settings to manage my
capacity at this movement. I am keen to know is there any other way to
display the actual CPU socket count on the ACS infrastructure dashboard ?

Thanks
AR

On 1 January 2015 at 00:38, Kevin McCormick <Ke...@intrinium.com>
wrote:

> I believe this is an issue with NUMA. I think this is only a display issue
> and not actually causing any capacity issues.
>
> We have this output from virsh nodeinfo on two identical hosts:
> CPU model:           x86_64
> CPU(s):              16
> CPU frequency:       1600 MHz
> CPU socket(s):       2 <--
> Core(s) per socket:  4
> Thread(s) per core:  2
> NUMA cell(s):        1 <--
> Memory size:         74225672 KiB
>
> CPU model:           x86_64
> CPU(s):              16
> CPU frequency:       1600 MHz
> CPU socket(s):       1 <--
> Core(s) per socket:  4
> Thread(s) per core:  2
> NUMA cell(s):        2 <--
> Memory size:         74225528 KiB
>
> The first host is configured with NUMA interleave mode in the system BIOS,
> which we've found as a better fit for our workload. That makes it appear as
> only a single NUMA cell with two sockets. The second system hasn't been set
> to interleave mode, and it's able to see both NUMA cells, each of which has
> a single CPU socket.
>
> Note that in both cases the total number of CPUs is detected correctly,
> and I believe that number is what ACS uses for capacity planning.
>
> Abdul, you may want to look into the cpu.overprovisioning.factor setting.
> This is at a cluster-level and affects how much CPU is reserved for an
> instance.
>
>
>
>
> Kevin McCormick
> Director of Cloud Services
>
>
> Intrinium
> Tel: (509) 465-1234 x7080
> Fax: (866) 565-4578
> Lync / Skype: Kevin.McCormick@intrinium.com
> Web: www.intrinium.com
> -----Original Message-----
> From: Nux! [mailto:nux@li.nux.ro]
> Sent: Wednesday, December 31, 2014 03:55
> To: users@cloudstack.apache.org
> Subject: Re: CPU socket miss match !!! Urgent
>
> Yes, I'm also seeing this on DELL PE R720... It seems like a libvirt fail,
> try to submit a bug upstream (redhat bugzilla).
>
> Lucian
>
> --
> Sent from the Delta quadrant using Borg technology!
>
> Nux!
> www.nux.ro
>
> ----- Original Message -----
> > From: "Abdul Rasool" <ra...@gmail.com>
> > To: users@cloudstack.apache.org
> > Sent: Wednesday, 31 December, 2014 07:49:37
> > Subject: Re: CPU socket miss match !!! Urgent
>
> > Hi Harikrishna,
> >
> > Thanks for your reply
> >
> > Here is the command out put.
> >
> > root@cloud-node001:~# dmidecode -t 4 |grep Socket
> > Socket Designation: CPU1
> > Upgrade: Socket LGA2011
> > Socket Designation: CPU2
> > Upgrade: Socket LGA2011
> >
> >
> > root@cloud-node001:~# virsh nodeinfo
> > CPU model:           x86_64
> > CPU(s):              8
> > CPU frequency:       2500 MHz
> > CPU socket(s):       1
> > Core(s) per socket:  4
> > Thread(s) per core:  1
> > NUMA cell(s):        2
> > Memory size:         32899468 kB
> >
> > When I check using dmidecode I am getting actual CPU sockets =2 whereas
> > libvirt is picking only one socket. please help me to fix this issue.
> >
> > Thanks
> > AR
> >
> >
> >
> > On 31 December 2014 at 11:27, Harikrishna Patnala <
> > harikrishna.patnala@citrix.com> wrote:
> >
> >> Hi,
> >>
> >> This seems to me like a libvirt issue on KVM.
> >> We read the socket count from the parameter ‘sockets’ in NodeInfo on KVM
> >> host.
> >> Can you check with the command “virsh nodeinfo” ?
> >>
> >> Thanks,
> >> Harikrishna
> >>
> >>
> >> On 31-Dec-2014, at 9:54 am, Abdul Rasool <rasoolsam@gmail.com<mailto:
> >> rasoolsam@gmail.com>> wrote:
> >>
> >> Attachment is missing ... added the same.
> >>
> >> On 31 December 2014 at 09:50, Abdul Rasool <rasoolsam@gmail.com<mailto:
> >> rasoolsam@gmail.com>> wrote:
> >> Hi,
> >>
> >> In my test / dev environment I have total 2 Dell Poweredge M620 blade
> >> servers configured as KVM hypervisor each has physical CPU of 2 sockets,
> >> total number of CPU sockets available in my setup is 4 sockets. but in
> ACS
> >> infrastructure shows only 2 physical sockets. for your reference I have
> >> attached my infrastructure screenshot.
> >>
> >> could you please some one help me why ACS is showing only 2 CPU sockets
> >> instead of 4 ? is there any configuration issue from my side. ? right
> now
> >> this environment is used for our internal development purpose due to CPU
> >> mismatch I am not able to add additional guest VM in my infrastructure
> it
> >> is throwing me insufficient resources.......!
> >>
> >> [Inline images 1]
> >>
> >> Regards,
> >> AR
> >>
> >>
> >>
>
>
> This email and any files transmitted with it are confidential and intended
> solely for the use of the individual or entity to whom they are addressed.
> If you have received this email in error please notify the system manager.
> This message contains confidential information and is intended only for the
> individual named. If you are not the named addressee you should not
> disseminate, distribute or copy this e-mail. Please notify the sender
> immediately by e-mail if you have received this e-mail by mistake and
> delete this e-mail from your system. If you are not the intended recipient
> you are notified that disclosing, copying, distributing or taking any
> action in reliance on the contents of this information is strictly
> prohibited.
>
>