Open Stack

> Yes, this is still happening. Mea culpa for not carrying the ball and
> maintaining visibility. There's work in nova to actually get it
> working, and in intel-nfv-ci to lay down the groundwork for eventual
> CI.
>
> In nova, the spec has been re-proposed for Stein [1]. There are some
> differences from the Rocky version, but based on what I've heard was
> discussed at Denver, it shouldn't be too controversial. There's a
> couple of nova patches up as well [2], but that's still pretty WIP
> given the changes in the spec. A bunch of patches from Rocky were
> abandoned because they're no longer applicable.
>
> In intel-nfv-ci, there's a whole stack of changes [3] that are mostly
> about technical debt and laying the groundwork to support multinode
> test environments, but there's also a WIP patch in there [4] that'll
> eventually actually test live migration.
>
> For now we have no upstream/public environment to run that on, so
> anyone who's involved will need their own env if they want to run the
> tests and/or play with the feature. Longer-term, I would like to have
> some form of upstream CI testing this, be it in the vanilla nodepool
> with nested virt and "fake" NUMA topologies, or a 3rd party CI with
> resources provided by an interested stakeholder.

Forgot the nova tag :(

> [1] https://review.openstack.org/#/c/599587/
> [2] https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:bp/numa-aware-live-migration
> [3] https://review.openstack.org/#/c/576602/
> [4] https://review.openstack.org/#/c/574871/6



-- 
--
Artom Lifshitz
Software Engineer, OpenStack Compute DFG