[nova][stable] Adding nova-core to nova stable branches

Előd Illés elod.illes at est.tech
Tue Jul 13 14:59:26 UTC 2021


Hi,

I still agree with what melwitt wrote on Victoria PTG etherpad and still 
keep my opinion what I wrote at that time: the nova-core and 
nova-stable-maint are quite overlapping, so why not just keep them 
separate and add the 2-4 active nova-core members to nova-stable-maint 
IF THEY WANT to deal with stable backports? To be clear, if more 
experienced stable cores agree that nova-core should be added to 
nova-stable-maint, then I'm not against it, though... It's just my opinion.

Előd


On 2021. 06. 25. 13:10, Stephen Finucane wrote:
> Happy Friday (again),
>
> Following on from my earlier email regarding os-vif stable branch permissions,
> I'd like to once again propose that we add nova-core to the list of teams with
> +2 rights on nova stable branches. Unlike the os-vif change, this has been
> discussed extensively in the past, most recently during the Victoria PTG [1]
> (search for "Can we add nova-core to nova-stable-maint") and therefore carries a
> little more baggage with it.
>
> Last time we discussed this, there were concerns from some existing stable cores
> that not all of nova-core were sufficiently well acquainted with stable branch
> policy and it was suggested that review velocity was sufficient. However, the
> stable team has remained pretty stable (heh) over the last couple of cycles and
> is staffed with developers with ample experience of both nova itself and the
> backport policies of nova and OpenStack as a whole. In addition, we continue to
> have a long tail of open stable reviews [2] from various people. The stable
> branches are increasingly important for users and organizations alike, and a
> faster, smoother backport policy encourages people to have an upstream-first
> mentality when it comes to maintaining a product built on these branches. The
> alternative is to handle these backports in a downstream-first or downstream-
> only manner, which results in duplication of effort and, over the long-term,
> potential difficulties keeping downstream and upstream in-sync.
>
> I think the time has long since passed where we should remove the artificial
> barrier that exists here. I expect all nova cores will have the maturity and
> sensibility to ask questions where uncertainty exists and only approve backports
> that they are confident and knowledgeable about. I do not expect existing stable
> cores will be sidelined either. Far from it, in fact. Existing stable cores will
> continue to be valuable both for their reviews and knowledge of the intricacies
> of stable policy.
>
> I proposed a waiting period of one week before making changes to os-vif
> permissions, but given the history around changes to nova stable core, I think
> we would benefit from some additional time to assuage any concerns that may
> still exist. I propose that we come to a decision by 09-Aug, or two weeks from
> now. If necessary, we can discuss this in real-time during the nova team meeting
> next Tuesday.
>
> Cheers,
> Stephen
>
> [1] https://etherpad.opendev.org/p/nova-victoria-ptg
> [2] https://review.opendev.org/q/project:openstack/nova+NOT+branch:master+is:open
>
>




More information about the openstack-discuss mailing list