View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0016025||CentOS-7||qemu-kvm||public||2019-04-22 12:41||2019-04-22 23:05|
|Target Version||Fixed in Version|
|Summary||0016025: BZ1638835 - KVM Windows 10 1803+ high cpu usage|
|Description||When using Centos 7.6 1810 as a VM host for Windows 10 1803 or higher there is very high idle cpu usage. On my low powered virtualisation host idle CPU usage is 150-200%.|
|Steps To Reproduce||Install CentOS 7.6 1810, install KVM virtualisation. install Windos 7 1803 or later guest. Observe idle CPU usage is high.|
|Additional Information||Fixed upstream in January: https://access.redhat.com/errata/RHBA-2019:0209|
|Tags||No tags attached.|
The fix you point to is not for the version of kvm included in the base/updates repos. It's for RHEV which is a different product. You will need to report it on bugzilla as affecting the version of qemu-kvm that we ship as part of the distro if you want that fixed.
We do also ship the qemu-kvm-ev package that is mentioned there as part of centos-release-qemu-ev.noarch
qemu-kvm.x86_64 10:1.5.3-160.el7_6.1 updates
is the distro version of qemu-kvm. Is that the one you're running?
Yes qemu-kvm.x86_64 10:1.5.3-160.el7_6.1 updates is the version I am on.
Is there anything special needed other than installing centos-release-qemu-ev.noarch in order to switch to the improved version?
yum list qemu\*-ev
I think they replace the existing ones, not sure if you have to remove the originals first.
Thanks, the process I eneded up following was:
yum group remove Virtualization\ Host
yum install centos-release-qemu-ev
yum install qemu-kvm-ev
virsh edit win10
to the <hyperv> section. Shut down / started the VM. There was some other mess involved that was specific to a mistake I made that broke the guest. But the issue appears to be resolved.
Many thanks, im happy for this post to be made public as I couldnt find these steps online at all and had earlier assumed that the resource usage couldnt be fixed on CentOS 7 without some updates from you folks.