fluo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Christopher <ctubb...@apache.org>
Subject Re: [DISCUSS] Relax RTC for urgent website updates
Date Thu, 27 Jul 2017 19:04:58 GMT
How about something like:

Reviews may only be bypassed in the case of an emergency, and only after an
appropriate level of effort has been made to request a review. Before
bypassing a review, the committer should take appropriate steps to ensure
that their actions can be reversed, if necessary. The committer should also
send an explanation to the dev list (or private list, if sensitive)
immediately afterwards justifying the urgency. The PMC members will decide
whether the action was warranted, and what follow-on actions should be
taken, if any. Such emergencies should be extremely rare, and would
typically only apply to things not blocked by a release vote (such as a
website change).

The above proposed wording has the benefit of being simple, flexible, and
accountable, but not being tied to any complex rules like tagging,
labeling, waiting for specific durations, etc. that can undermine the
efficacy of an emergency action. It's also very broad, so it isn't
dependent on specific workflows or infrastructure tools, which can change
over time.

On Thu, Jul 27, 2017 at 11:12 AM Keith Turner <keith@deenlo.com> wrote:

> In addition to tagging as urgent, a short explanation of why its
> urgent should be given.
> On Thu, Jul 27, 2017 at 11:10 AM, Keith Turner <keith@deenlo.com> wrote:
> > I have been thinking that it may be useful to relax RTC for urgent
> > website updates.   I can not imagine this being needed for Fluo or
> > Fluo Recipes because of the 3 day release process.  However, the
> > website is always immediately available after any update.  It would be
> > nice to have an agreed on mechanism for bypassing RTC for the website.
> > Possibly something like the following :
> >
> >  * Create PR for website and tag it urgent
> >  * Attempt to contact other PMC members
> >  * Wait X time (for example 10 mins)
> >  * After X time if no one has indicated they are reviewing, then
> > commiter can push
> >
> > I think the policy should include something like : should this policy
> > ever cause strife in the community, it must be repealed immediately.

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message