[openstack-dev] [nova] Announcing my candidacy for PTL

Matt Riedemann mriedem at linux.vnet.ibm.com
Mon Mar 14 18:46:21 UTC 2016


Hi everyone, this is my announcement to run for Nova PTL in the Newton 
cycle.

First off, I see the Nova PTL role as a service position to the rest of 
the team. The job of the PTL is to organize the release schedule, help 
people coordinate and do a lot of the administrative work with the 
release team and other project teams to try and keep things as smooth as 
possible for developers to do what they do best, with as little 
distraction as possible. Given Nova's maturity and existing strong 
technical development team, I see the PTL now as less of a technical 
overlord and as more of a manager, and that's how I'd try to work.

I've noticed more sub-teams organizing within Nova in the last cycle or 
two and really want to continue to foster that model. For example, there 
were dedicated sub-teams around priorities like cells v2, live 
migration, notifications and the scheduler. Those sub-teams were holding 
regular meetings and bringing forth issues and status to the broader 
team as necessary. I think this makes those priority efforts more 
productive since we work better when focused for shorter
periods of time on specific goals.

John did a great job of fostering others to own parts of the work via 
CPLs, the bug triage team, and bringing up potential nova core 
additions. I want to continue that work since I think having different 
people in these roles helps them grow within the project and the 
community. I'd like to also get Nova more involved with the operator 
community, which probably starts with a CPL attending the operator 
meeting and bringing back news to the Nova team. It amazes me how much 
we still generally work in silos and don't effectively community issues 
across groups of teams and I want to make that better.

Which brings me to my main focus for Newton, which is cross-project 
priorities. There are several items that need to make progress in Newton 
to move not only Nova but other projects forward. These include:

* Get-me-a-network (Nova/Neutron)
* os-vif integration (Nova/Neutron)
* os-brick + privsep integration (Nova/Cinder)
* Volume multiattach (Nova/Cinder)
* Glance v2 integration (Nova/Glance)

We also have to continue moving forward with Cells V2 and the scheduler 
/ resource tracker work. We already have sub-teams/owners for a few of 
these items but some are going to require renewed focus, and as PTL I'd 
be working with other projects on setting up plans/goals on how to 
accomplish these in Newton and then communicating those back to the 
broader team so we can all be on the same page. I loved seeing the 
status updates in the mailing list toward the end of Mitaka and want to 
continue seeing that type of communication.

We need to do _something_ about the release scheduling for blueprints, 
meaning priority and non-priority, and also having time at the end to 
stabilize and get bugs fixed. We spend a lot of time reviewing specs 
(like most of the first milestone), then non-priority features, and 
finally trying to sprint on the things we said were priorities but we 
for some reason schedule to work on at the end of the release, when 
every is tired and time is short to make big changes if necessary. I 
think we need to flip that somehow so that priorities actually get 
treated like priorities early in the cycle and have focused effort
by the core team.

And last but not least, QA is always a special focus of mine because if 
the gate is busted or we're landing busted code for a lack of testing, 
it doesn't really matter what else we do. That's why we need to get a 
Ceph job back voting reliably on Nova (and Cinder/Glance), along with a 
trunk libvirt job we've been talking about since Kilo - which will also 
feed into a dedicated live migration job. So as PTL those would be 
things I'd want to help move forward.

If you've made it through all of this, thanks and congratulations! 
Regardless of the election I plan to do my best to serve the project and 
the people working on it in Newton.

Thanks,

Matt Riedemann




More information about the OpenStack-dev mailing list