If you are deploying on CentOS (with RDO?), you can enable the CentOS Virtualization special interest group [1] repository.
The repository contains qemu-kvm-ev >= 2.3 backported from RHEV. It is recommended as the qemu-kvm version from base CentOS repositories is not high enough and lacks some features (things like snapshots, iirc). qemu-kvm >= 2.3 is actually a requirement in RDO >= Newton and we'll bundle the CentOS virtualization SIG repository in our release packages. [1]: https://wiki.centos.org/SpecialInterestGroup/Virtualization David Moreau Simard Senior Software Engineer | Openstack RDO dmsimard = [irc, github, twitter] On Thu, Sep 29, 2016 at 5:00 AM, William Josefsson <william.josef...@gmail.com> wrote: > thanks everyone, I verified setting mem_stats_period_seconds = 0 as > suggested by Corbin in nova.conf libvirt section, and then restarting > openstack-nova-compute service and it works! > > While this seems to be a workable workaround I'm not sure what's the plans > to permanently fix this in CentOS7.2? thx will > > > > On Wed, Sep 28, 2016 at 11:37 PM, Corbin Hendrickson > <corbin.hendrick...@endurance.com> wrote: >> >> Oh you can read it in the bug thread, but I forgot to mention, if you put >> in your nova.conf under the libvirt section mem_stats_period_seconds = 0, >> and restart nova on the destination (although i'd say just do it on both) it >> will no longer hit the bug. I tested this a couple weeks back with success. >> >> Corbin Hendrickson >> Endurance Cloud Development Lead - Manager >> Cell: 801-400-0464 >> >> On Wed, Sep 28, 2016 at 9:34 AM, Corbin Hendrickson >> <corbin.hendrick...@endurance.com> wrote: >>> >>> It unfortunately is affecting virtually all of Redhat's latest qemu-kvm >>> packages. The bug that was unintentionally introduced was done so in >>> response to CVE-2016-5403 Qemu: virtio: unbounded memory allocation on host >>> via guest leading to DoS. >>> >>> Late in the bug thread, they finally posted to a new bug created for the >>> breaking of live migrate via Bug 1371943 - RHSA-2016-1756 breaks migration >>> of instances. >>> >>> Based off their posts i've been following it's likely going to "hit the >>> shelves" when RHEL 7.3 / CentOS 7.3 comes out. It does look like they are >>> backporting it to all their versions of RHEL so that's good. >>> >>> But yes this does affect 2.3 as well. >>> >>> Corbin Hendrickson >>> Endurance Cloud Development Lead - Manager >>> Cell: 801-400-0464 >>> >>> On Wed, Sep 28, 2016 at 9:13 AM, Van Leeuwen, Robert >>> <rovanleeu...@ebay.com> wrote: >>>> >>>> > There is a bug in the following: >>>> >>>> > >>>> >>>> > qemu-kvm-1.5.3-105.el7_2.7 >>>> >>>> > qemu-img-1.5.3-105.el7_2.7 >>>> >>>> > qemu-kvm-common-1.5.3-105.el7_2.7 >>>> >>>> >>>> >>>> You might be better of using the RHEV qemu packages >>>> >>>> They are more recent (2.3) and have more features compiled into them. >>>> >>>> >>>> >>>> Cheers, >>>> >>>> Robert van Leeuwen >>>> >>>> >>>> _______________________________________________ >>>> OpenStack-operators mailing list >>>> OpenStack-operators@lists.openstack.org >>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators >>>> >>> >> > > > _______________________________________________ > OpenStack-operators mailing list > OpenStack-operators@lists.openstack.org > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators > _______________________________________________ OpenStack-operators mailing list OpenStack-operators@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators