[openstack-dev] Fixing the Nova Core Reviewer Frustration
James.Bottomley at HansenPartnership.com
Tue Apr 7 17:58:26 UTC 2015
On Tue, 2015-04-07 at 13:35 -0400, Anita Kuno wrote:
> On 04/07/2015 01:02 PM, James Bottomley wrote:
> > On Tue, 2015-04-07 at 11:27 +1000, Michael Still wrote:
> >> Additionally, we have consistently asked for non-cores to help cover
> >> the review load. It doesn't have to be a core that notices a problem
> >> with a patch -- anyone can do that. There are many people who do help
> >> out with non-core reviews, and I am thankful for all of them. However,
> >> I keep meeting people who complain about review delays, but who don't
> >> have a history of reviewing themselves. That's confusing and
> >> frustrating to me.
> > I can understand why you're frustrated, but not why you're surprised:
> > the process needs to be different. Right now the statement is that for
> > a patch series to be accepted it has to have a positive review from a
> > core plus one other, however the "one other" can be a colleague, so it's
> > easy. The problem, as far as submitters see it, is getting that Core
> > Reviewer. That's why so much frenzy (which contributes to your
> > frustration) goes into it. And why all the complaining which annoys
> > you.
> > To fix the frustration, you need to fix the process: Make the cores
> > more of a second level approver rather than a front line reviewer and I
> > predict the frenzy to get a core will go down and so will core
> > frustration. Why not require a +1 from one (or even more than one)
> > independent (for some useful value of independent) reviewer before the
> > cores will even look at it? That way the cores know someone already
> > thought the patch was good, so they're no longer being pestered to
> > review any old thing and the first job of a submitter becomes to find an
> > independent reviewer rather than go bother a core.
> > James
> > __________________________________________________________________________
> > OpenStack Development Mailing List (not for usage questions)
> > Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> Hi James:
> Since this is now an open thread and no longer has anything to do with
> Nova PTL candidacy or anyone else's PTL candidacy, I'm going to jump in
> here with a recommendation.
> Since you are familiar with Gerrit yourself and have a merged patch,
> perhaps you can spend some time between now and summit and review as
> many Nova (or the project of your choice) patches as you can to learn
> what life is like from the reviewers point of view.
Thanks for the suggestion. However, I didn't make my initial
recommendation based on not having some insight into what's going on.
I'm SCSI Subsystem maintainer of Linux, meaning I act like an OpenStack
core for all the patches that go into this subsystem.
We recently hit a crisis point in Linux last year where I simply
couldn't review all the patches and someone else took over to help out.
He instituted a process whereby no patch got consideration until it had
at least one other review and even though he's stepped back again, I
find that adhering to this process brings my workload back to being
manageable because I can just tell anyone bothering me about a patch to
go away and find a reviewer. Once it has a reviewer (provided I trust
them), I merely need to glance over it to verify no problems before
including it in the git tree.
I'm basing my recommendation directly on how this process has helped me
continue to do my job in Linux.
Now, I think it's fair game to argue about whether this would, or would
not be applicable to OpenStack and whether the benefits we saw in Linux
would be fully realized in a different environment. I do, though, think
it's slightly unwise to dismiss out of hand experience gained in other
projects, unless you truly believe OpenStack has nothing to learn from
> If you find it supportive to do so please help yourself to this blog
> post I wrote about reviewing an OpenStack patch:
> Thanks James,
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
More information about the OpenStack-dev