Red Hat Bugzilla – Bug 828023
[libvirt] Setting numa parameters causes guest xml error
Last modified: 2013-02-21 02:16:14 EST
Description of problem: 'virDomain{G, S}etNumaParameters' APIs generate wrong guest xml Version-Release number of selected component (if applicable): libvirt-0.9.10-21.el6.x86_64 rhel6.3 rc build How reproducible: always Steps to Reproduce: 1.# python Python 2.6.6 (r266:84292, May 1 2012, 13:52:17) [GCC 4.4.6 20110731 (Red Hat 4.4.6-3)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import libvirt >>> conn = libvirt.open(None) >>> dom = conn.lookupByName("rhel6.3rc") >>> dom.numaParameters(0) {'numa_nodeset': '1', 'numa_mode': 1} >>> param = {'numa_nodeset': '1', 'numa_mode': 1} >>> dom.setNumaParameters(param, 2) 0 # destroy and start the domain in another console Actual results: the guest can't restart # virsh start rhel6.3rc error: Failed to start domain rhel6.3rc error: (domain_definition):8: error parsing attribute name <memory mode='preferred' </numatune> -------------------------------^ #virsh dumpxml rhel6.3rc ............. <vcpu placement='static'>1</vcpu> <numatune> <memory mode='preferred' </numatune> ............. Expected results: the guest can start without error Additional info: use command : #virsh numatune rhel6.3rcqcow2 --nodeset 1 --config will get the same error. refer bug :https://bugzilla.redhat.com/show_bug.cgi?id=771591 , this bug verify in libvirt-0.9.10-1, i can't reproduce this on libvirt-0.9.10-16.el6.x86_64 so it's a regression bug.
This problem hasn't any relationship with libvirt client such as virsh or python binding, it's a underlying implementatin issue, which probaly is introduced by upstream commit 8d26a7f or 97010eb1, however, I haven't check it further. Osier, could you help confirm this? thanks.
*** Bug 838049 has been marked as a duplicate of this bug. ***
v2 posted upstream. https://www.redhat.com/archives/libvir-list/2012-June/msg01018.html
commit 6fd1708fad221956c1ca7de023a7babfb100a4dc Author: Osier Yang <jyang@redhat.com> Date: Tue Aug 28 20:13:31 2012 +0800 qemu: Set placement when setting numa parameters To keep the internal data structure consistent.
pkgs: libvirt-0.10.0-1.el6.x86_64 qemu-kvm-rhev-0.12.1.2-2.297.el6_3.x86_64 kernel-2.6.32-279.el6.x86_64 steps: 1. # virsh start libvirt_test_api Domain libvirt_test_api started # virsh list --all Id Name State ---------------------------------------------------- 4 libvirt_test_api running 2. numatune # virsh numatune libvirt_test_api --nodeset 2 --config success, so no xml problem now. 3. destroy domain # virsh destroy libvirt_test_api Domain libvirt_test_api destroyed 4. start domain # virsh start libvirt_test_api Domain libvirt_test_api started # virsh dumpxml libvirt_test_api ... <numatune> <memory mode='strict' nodeset='2'/> </numatune> ... 5. check with python # python Python 2.6.6 (r266:84292, May 1 2012, 13:52:17) [GCC 4.4.6 20110731 (Red Hat 4.4.6-3)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import libvirt >>> conn = libvirt.open(None) >>> dom = conn.lookupByName("libvirt_test_api") >>> dom.numaParameters(0) {'numa_nodeset': '2', 'numa_mode': 0} >>> param = {'numa_nodeset': '4', 'numa_mode': 1} >>> dom.setNumaParameters(param, 2) 0 >>> # virsh destroy libvirt_test_api Domain libvirt_test_api destroyed # virsh start libvirt_test_api Domain libvirt_test_api started # virsh dumpxml libvirt_test_api ... <numatune> <memory mode='preferred' nodeset='4'/> </numatune> ...
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. http://rhn.redhat.com/errata/RHSA-2013-0276.html