On 09/15/2016 10:30 AM, Matthew Thode wrote: > On 09/15/2016 09:09 AM, Matt Riedemann wrote: >> There was a change introduced to nova in newton which requires a minimum >> of netaddr 0.7.13. We've been testing with 0.7.18 from upper-constraints >> so bumping the minimum from 0.7.12 to 0.7.13 shouldn't cause any issues. >> >> The g-r change is here: >> >> https://review.openstack.org/#/c/370833/ >> >> We're not going to revert the nova change since that would be riskier at >> this point in the cycle since it landed before n-2. >> > > This is a hard ask and I can't really support it, mainly because of all > the things it impacts this late in the cycle. Is there ANY way we can > not merge this? > > just some of the things that would need to be re-released > > > networking-ovn (just had their rc1) > os-net-config > os-vif > oslo.config > oslo.serialization > oslo.utils > oslo.versionedobjects > oslo.vmware > python-neutronclient > wsme
Actually, there is very little this impacts. Basically no systems use the g-r minimum except as a code comment. Yes, it will produce a patch to be merged on those projects, however the only project that needs to merge in is Nova, it's the only project that uses the function that will actually explode at that level. Given this is the lowest risk way to address this issue, far less risky than a barely tested Nova workaround, this is really the way things should move forward. Nothing is harmed in the release if there is a mix of 0.7.12 / 0.7.13 minimums. Doug is even proposing that we explicitly allow situations like that for the next cycle. Projects can merge and catch up in stable/newton post release if we feel consistency is important. -Sean -- Sean Dague http://dague.net __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev