Hi all, I know that XenSource allows us to configure how much memory and how many vcpus are available to the Domain0 through kernel boot parameters (dom0_mem and dom0_max_vcpus). XCP is underlying using XenSource, then these settings should also work with XCP.
I'm trying to specify these boot parameters in the /boot/extlinux.conf file of my Domain0 VMs (Control Domains): # /boot/exlinux.conf # ... label xe # XenServer append /boot/xen.gz dom0_max_vcpus=4 dom0_mem=752M lowmem_emergency_pool=16M crashkernel=...@32m ... The dom0_mem works just fine. My problem is with the dom0_max_vcpus boot parameter, which apparently isn't working. After setting this boot parameter, I also tried to set the number of vcpus of my Control Domains with xapi: xe vm-param-set VCPUs-at-startup=4 uuid=<ControlDomainUUID> xe vm-param-set VCPUs-max=4 uuid=<ControlDomainUUID> None of them worked. My domain0 is still seeing only one vcpu. The strange thing is that XenSource's dom0_mem boot parameter did work as expected, but domain0_max_vcpus didn't. Does someone know if XCP/XenServer is somewhere locking the number of Domain0 vCPUs to only one? Should I try something else? My main issue is that I have a (very) large number of Networks and VLANs active inside my XCP pools. And because of that, the xapi process in all hosts (control domains) constantly hits 100% of cpu usage and all XCP operations are very slow. Pools may take something close to 3 hours just to boot, because XCP hosts recreate and synchronize all networks and vlans right after boot. Setting the cpu priority of my control domains to higher values already helped, but I still have many unused cores in my physical machines. Being able to use them in my busy Domain0s would be very good. Cheers, -- Fabio Kung
_______________________________________________ xen-api mailing list [email protected] http://lists.xensource.com/mailman/listinfo/xen-api
