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

Re: [libvirt] [PATCH] nodeinfo: Add workaround if kernel reports bogous numa topology.



On Wed, Oct 31, 2012 at 4:50 PM, George-Cristian Bîrzan <gc birzan org> wrote:
on 0.10.2:

[root host34 libvirt]# virsh nodeinfo
CPU model:           x86_64
CPU(s):              24
CPU frequency:       2200 MHz
CPU socket(s):       1
Core(s) per socket:  6
Thread(s) per core:  1
NUMA cell(s):        4
Memory size:         131971020 KiB

and http://birzan.org/capabilities4-newlibvirt.txt 

This has made the problem even worse, as we now can only use 6 cores out of 24 by default (libvirt pins qemus to the CPUs is sees available, so we have to manually taskset them after starting).


I should check before I let my mouth run. This actually fixed the problem, we now get all 24 CPUs:

virsh # vcpuinfo 08867271-d458-43dc-8a03-807e83d107b4
VCPU:           0
CPU:            9
State:          running
CPU time:       1.3s
CPU Affinity:   yyyyyyyyyyyyyyyyyyyyyyyy


To give you an example from the 'broken' host (still with 0.10.2):

[root host29 ~]# virsh vcpuinfo 17db9ebb-691b-47e1-bbf4-c9e3c1388d4f
VCPU:           0
CPU:            3
State:          running
CPU time:       3103.2s
CPU Affinity:   yyyyyyyyyyyy

VCPU:           1
CPU:            1
State:          running
CPU time:       3460.4s
CPU Affinity:   yyyyyyyyyyyy
 

--
George-Cristian Bîrzan

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