Hide Forgot
Description of problem: when there is no numatune for memory, 'auto' placement vcpu will add 'auto' placement memory numatune automatically. But user specified memnode for each guest NUMA node in numatune, when 'auto' placement vcpu without memory placement. The libvirtd will add memory placement to 'auto' and remove memnode. Although this configuration doesn't align vcpus, but it should be valid. Version-Release number of selected component (if applicable): libvirt-1.2.8-15.el7.x86_64 qemu-kvm-rhev-2.1.2-21.el7.x86_64 3.10.0-223.el7.x86_64 How reproducible: 100% Steps to Reproduce: 1. prepare guest xml # cat dummy.xml <domain type='qemu'> <name>dummy</name> <memory unit='KiB'>2097152</memory> <vcpu placement='auto'>4</vcpu> <numatune> <memnode cellid="0" mode="strict" nodeset="1"/> <memnode cellid="1" mode="preferred" nodeset="0"/> </numatune> <os> <type arch='x86_64' machine='pc'>hvm</type> </os> <cpu> <numa> <cell id='0' cpus='0-1' memory='1048576'/> <cell id='1' cpus='2-3' memory='1048576'/> </numa> </cpu> </domain> # virsh define dummy.xml 2. then check xml, the <memnode> is lost # virsh dumpxml dummy <domain type='qemu'> <name>dummy</name> <uuid>1fd33000-27cb-425c-94f9-c7454914acdb</uuid> <memory unit='KiB'>2097152</memory> <currentMemory unit='KiB'>2097152</currentMemory> <vcpu placement='auto'>4</vcpu> <numatune> <memory mode='strict' placement='auto'/> </numatune> <os> <type arch='x86_64' machine='pc-i440fx-rhel7.1.0'>hvm</type> <boot dev='hd'/> </os> <cpu> <numa> <cell id='0' cpus='0-1' memory='1048576'/> <cell id='1' cpus='2-3' memory='1048576'/> </numa> </cpu> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <emulator>/usr/libexec/qemu-kvm</emulator> <controller type='usb' index='0'> <address type='pci' domain='0x0000' bus='0x00' slot='0x01' function='0x2'/> </controller> <controller type='pci' index='0' model='pci-root'/> <memballoon model='virtio'> <address type='pci' domain='0x0000' bus='0x00' slot='0x02' function='0x0'/> </memballoon> </devices> </domain> Expect result: memnode is an other memory placement configuration, 'auto' vcpu placement should not remove it.
Fixed upstream: commit 638e3d270fc4b1ebeab0d139c2b4e54f590eb8b5 Author: Peter Krempa <pkrempa> Date: Thu Feb 12 16:04:45 2015 +0100 conf: numa: Refactor logic in virDomainNumatuneParseXML Shuffling around the logic will allow to simplify the code quite a bit. As an additional bonus the change in the logic now reports an error if automatic placement is selected and individual placement is configured. v1.2.12-174-g638e3d2
I can reproduce this issue with libvirt-1.2.8-16.el7.x86_64: 1. add <memnode> in numatune settings: # virsh edit test3 ... <vcpu placement='auto'>4</vcpu> <numatune> <memnode cellid="0" mode="strict" nodeset="1"/> <memnode cellid="1" mode="preferred" nodeset="0"/> </numatune> ... 2. recheck the xml: # virsh dumpxml test3 ... <vcpu placement='auto' current='1'>2</vcpu> <numatune> <memory mode='strict' placement='auto'/> </numatune> ... And verify this bug with libvirt-1.2.15-2.el7.x86_64: 1. try to add <memnode> in numatune settings and set numa placement is auto: # virsh edit test3 <vcpu placement='auto'>4</vcpu> <numatune> <memnode cellid="0" mode="strict" nodeset="1"/> <memnode cellid="1" mode="preferred" nodeset="0"/> </numatune> error: unsupported configuration: Per-node binding is not compatible with automatic NUMA placement. Failed. Try again? [y,n,i,f,?]:
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2015-2202.html