Ok, thanks Andrey.

I experience a behavior that can be placed in your explanation (dough I 
certainly not a expert :-) ). If SQL guest has a CPU load of 5%, it takes up 
50% on the host. Sometimes (late in the evening), the gest cpu load drops to 
almoust 0% and the host CPU-load drops to about 10% (same as the freshly 
installed server). So maybe due to the non-activity of the SQL-clients, the 
SQL-server goes in a kind of rest-state and stops polling for a while until a 
client accesses the database again or a maintenance plan is started...

I'll settle for now and stop tweaking.

Here below, a summary of tips:
- Setting video to QXL and the display channel to Spice
        <video>
                 <model type='qxl' ram='65536' vram='65536' heads='1'/>
                <address type='pci' domain='0x0000' bus='0x00' slot='0x02' 
function='0x0'/>
        </video>
        <channel type='spicevmc'>
                <target type='virtio' name='com.redhat.spice.0'/>
                <address type='virtio-serial' controller='0' bus='0' port='1'/>
        </channel>
- Using HyperV enlightenemt timer
        <features>  
                <hyperv>  
                        <relaxed state='on'/>  
                        <vapic state='on'/>  
                        <spinlocks state='on' retries='8191'/>  
                </hyperv>  
         <features/>  
        <clock ...>  
                <timer name='hypervclock' present='yes'/>  
        </clock>
- Delete the tablet entry in the input section

Again thanks Daniel and Andrey.

-----Oorspronkelijk bericht-----
Van: Andrey Korolyov [mailto:and...@xdel.ru] 
Verzonden: vrijdag 13 maart 2015 20:16
Aan: Dominique Ramaekers
CC: Daniel P. Berrange; libvirt-users@redhat.com
Onderwerp: Re: [libvirt-users] Processor usage of qemu process.

As far as I can understand the mentioned improvement is targeted on a polling 
mode for a MSSQL, there is no wonder for observing relatively high hypervisor 
CPU consumption without any sign of same consumption in a guest itself. Certain 
kinds of applications (PBXes, Chrome/Chromium, seemingly MSSQL) may poll some 
resource very frequently, causing wakeups in the host for guest process and 
increasing overall running time of a guest process. I`ve heard of none of the 
mitigations for this class of workload issues, though you can experiment with 
guest timer settings if you want to. As we are talking in the libvir-list, it 
is safe to suggest to use native HyperV host to eliminate this problem 
completely. :)

_______________________________________________
libvirt-users mailing list
libvirt-users@redhat.com
https://www.redhat.com/mailman/listinfo/libvirt-users

Reply via email to