Open Stack

Hi,

Getting so mixed that I’ll jump to the inline commenting as well.

From: Boris Pavlovic [mailto:boris at pavlovic.me]
Sent: 13 February 2015 15:01
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [glance] Cleanout of inactive change proposals from review

Erno,


My personal take is that if some piece of work has not been touched for a month, it’s probably not that important after all and the community should use the resources to do some work that has actual momentum.

Based on my experience, one of the most common situation in OpenStack is next:
1) Somebody makes fast (but with right idea) changes, because he (and usually others) needs it
2) It doesn't pass review process fast
3) Author of this patch has billions others tasks (not related to upstream) and
can't work on this change anymore
4) Patch get's abounded and forgotten

I’m unfortunately starting to sound like a broken record but again. If no-one has touched the change (or taken it over) in 4 weeks at the point when there is clear indication that if the change does not get traction it will be cleaned from the review, it’s probably not worth of keeping there any longer either.

The changes itself will not disappear the owner is still able to revive it if felt that there is right time to continue it.

Nobody never reviews abounded changes..

Repeating the previous, if your change gets abandoned because of inactivity and you don’t care about it, why should someone else who haven’t cared so far?

 The cleanup will just make it easier for people to focus on things that are actually moving.

Making decision based on activity around patches is not the best way to do things.

So what would be better way to do it? We have currently 4 pages of change proposals in the review that has been touched by anyone in Feb. Honest question, who scrolls further than that or even down to that 4th page? From page 6 forward there are changes that has been last time touched last year. And this is purely from “updated” column, so I did not look when the owner/author/committer has last time touched these.

If we take a look at the structure of OpenStack projects we will see next things:

1) Things that re moving fast/good are usually related to things that core team (or active members) are working on.
This team is resolving limited set of use cases (this is mostly related because not every member is running it's own production cloud)

This is very true, dropping that core team away and let’s keep the active members here. Because it’s a community it’s extremely difficult to get people working on something else than what they or their employers sees important.

2) Operators/Admins/DevOps that are running their own cloud have a lot of experience knows a lot of missing use cases and
source of issues. But usually they are not involved in community process so they don't know whole road map of project, so they are not able to fully align their patches with road map, or eventually just don't have enough time to work on features.

So abounding patches from 2 group just because of inactivity can make big harm to project.

I don’t think pushing for activity is bad thing and would do big harm for project(s). These are matters of priority and I do not see any benefit keeping changes in review that haven’t been touched for months (current situation). If this group 2 is the fundamental issue of our changes stalling in review, we need to fix that rather than let it clutter the queue. We are talking open source project and community here. I find it extremely hard to justify asking anyone in the community to take responsibility of someone else’s production cloud if they have no interest to resource the above for the benefit of their own business.

 Do you have resources in mind to dedicate for this?

Sometimes I am doing it by my self, sometimes newbies in community (that want some work to get involved), sometimes core team is working on old patches.

We will not run out of the bug fixing work and the commits against those bugs will stay in the bug even after they get abandoned.

Important chagesets are supposed to have bugs (or blueprints) assigned
to them, so, even if the CS is abandoned, its description still
remains on Launchpad in one form or another, so we will not loose it
from general project's backlog

This is not true in a lot of cases. =)
 In many cases DevOps/Operators don't know or don't want to spend time for launchpad/specs/ and so on.

Then we need to educate and encourage them instead of support the behavior of “Throw it in and someone at some day will maybe take care of it”-attitude.


-                  Erno

Best regards,
Boris Pavlovic


On Fri, Feb 13, 2015 at 5:17 PM, Kuvaja, Erno <kuvaja at hp.com<mailto:kuvaja at hp.com>> wrote:
Hi Boris,

Thanks for your input. I do like the idea of picking up the changes that have not been active. Do you have resources in mind to dedicate for this?

My personal take is that if some piece of work has not been touched for a month, it’s probably not that important after all and the community should use the resources to do some work that has actual momentum. The changes itself will not disappear the owner is still able to revive it if felt that there is right time to continue it. The cleanup will just make it easier for people to focus on things that are actually moving. It also will make bug tracking bit easier when one will see on the bug report that the patch got abandoned due to inactivity and indicates that the owner of that bug might not be working on it after all.


-          Erno

From: Boris Pavlovic [mailto:bpavlovic at mirantis.com<mailto:bpavlovic at mirantis.com>]
Sent: Friday, February 13, 2015 1:25 PM
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [glance] Cleanout of inactive change proposals from review

Hi,

I believe that keeping review queue clean is the great idea.
But I am not sure that set of these rules is enough to abandon patches.

Recently I wrote blogpost related to making OpenStack community more user friendly:
http://boris-42.me/thoughts-on-making-openstack-community-more-user-friendly/

tl;dr;

Patches on review are great source of information what is missing in project.
Removing them from queue means losing this essential information. The result
of such actions is that project doesn't face users requirements which is quite bad...

What if that project team continue work on all "abandoned" patches  that are covering
valid use cases and finish them?

Best regards,
Boris Pavlovic



On Fri, Feb 13, 2015 at 3:52 PM, Flavio Percoco <flavio at redhat.com<mailto:flavio at redhat.com>> wrote:
On 13/02/15 11:06 +0000, Kuvaja, Erno wrote:
Hi all,

We have almost year old (from last update) reviews still in the queue for
glance. The discussion was initiated on yesterday’s meeting for adopting
abandon policy for stale changes.

The documentation can be found from https://etherpad.openstack.org/p/
glance-cleanout-of-inactive-PS and any input would be appreciated. For your
convenience current state below:

Thanks for putting this together. I missed the meeting yday and this
is important.
Glance - Cleanout of inactive change proposals from review


We Should start cleaning out our review list to keep the focus on changes that
has momentum. Nova is currently abandoning change proposals that has been
inactive for 4 weeks.



Proposed action (if all of the following is True, abandon the PS):

1. The PS has -1/-2 (including Jenkins)

I assume you're talking about voting -1/-2 and not Workflow, right?
(you said jenkins afterall but just for the sake of clarity).
2. The change is proposed to glance, glance_store or python-glanceclient;
   specs should not be abandoned as their workflow is much slower

3. No activity for 28 days from Author/Owner after the -1/-2

I'd reword this in "No activity". This includes comments, feedback,
discussions and or other committers taking over a patch.
4. There has been  query made to the owner to update the patch between 5 and
   10 days  before abandoning (comment on PS/Bug or something similar)

 ● Let's be smart on this. Flexibility is good on holiday seasons, during
   feature freeze, etc.

+2 to the above, I like it.

Thanks again,
Flavio

--
@flaper87
Flavio Percoco

__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe<http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe<http://OpenStack-dev-request@lists.openstack.org?subject:unsubscribe>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20150213/f4b621d7/attachment.html>