Monday, 2018-02-12

*** gcb has joined #openstack-tc01:05
*** gcb has quit IRC01:18
*** gcb has joined #openstack-tc02:30
*** gcb has quit IRC03:07
*** gcb has joined #openstack-tc03:20
*** kumarmn has joined #openstack-tc03:21
*** kumarmn_ has quit IRC03:24
*** ykarel|away has joined #openstack-tc04:25
*** ykarel|away is now known as ykarel04:44
*** gcb has quit IRC06:41
*** gcb has joined #openstack-tc06:56
*** gcb has quit IRC07:02
*** eumel8 has joined #openstack-tc07:24
*** dtantsur|afk is now known as dtantsur08:33
*** bauwser is now known as bauzas08:43
*** jpich has joined #openstack-tc08:54
*** ykarel is now known as ykarel|lunch09:15
ttxthe talk voting, i think, achieves two objectives: (1) give a basic set of inputs to help track chairs select talks, and (2) generate noise/excitement around the conference09:21
ttxTalk voting however makes for a pretty bad input set. There was this proposed system of peer selection, giving each speaker a random set of n talks to rate, and use that instead.09:24
ttxhowever that fails objective (2) :)09:25
*** dirk has quit IRC09:33
*** dirk has joined #openstack-tc09:34
*** ykarel|lunch is now known as ykarel10:19
*** cdent has joined #openstack-tc10:32
*** mrhillsman has quit IRC10:57
*** mrhillsman has joined #openstack-tc10:58
*** rosmaita has joined #openstack-tc13:18
*** kumarmn_ has joined #openstack-tc13:29
*** kumarmn has quit IRC13:32
*** eumel8 has quit IRC13:48
ttxtc-members: I just sent an email to the -tc list about PTG postlunch presentations. Please read so that we can finalize the choice by tomorrow!13:51
*** dtantsur is now known as dtantsur|brb13:52
fungiand #2 is a double-edged sword with prospective speakers posting "vote for my talk" all over including in inappropriate forums where others quickly get annoyed and openstack ends up with a negative reputation over it13:54
mugsiefungi: ++13:54
*** kong has quit IRC13:55
persiaTrick is balancing the benefits of cross-community pollination vs. cross-community irritation.  Leaving it to speakers theoretically allows OpenStack to disavow the misbehaviour, but that may not overcome the distaste.13:56
*** kumarmn_ has quit IRC14:12
*** kumarmn has joined #openstack-tc14:30
*** david-lyle has quit IRC14:38
*** ykarel is now known as ykarel|away15:44
*** david-lyle has joined #openstack-tc15:46
*** ykarel|away has quit IRC15:54
*** ykarel|away has joined #openstack-tc16:07
*** ykarel|away has quit IRC16:16
*** dtantsur|brb is now known as dtantsur16:24
*** hongbin has joined #openstack-tc16:51
*** jpich has quit IRC17:07
*** ykarel|away has joined #openstack-tc17:53
*** david-lyle has quit IRC17:55
*** dklyle has joined #openstack-tc18:07
*** diablo_rojo has joined #openstack-tc18:09
openstackgerritThierry Carrez proposed openstack/governance master: Team diversity tags update, February 2018
*** harlowja has joined #openstack-tc18:14
*** dklyle has quit IRC18:30
*** david-lyle has joined #openstack-tc18:54
*** ykarel|away has quit IRC18:57
dmsimardOuch, glance almost becoming single vendor is worrying -- did the "top 5 help requested" ever translate into anything meaningful ?19:38
*** lbragstad_ has joined #openstack-tc19:41
cdentdmsimard: it depends on how you ask. What was discovered at the time was that the discussion surrounding its inclusion on the list was useful for drawing attention to the problem, but the list itself, not so much.19:42
*** lbragstad_ has quit IRC19:43
*** dtantsur is now known as dtantsur|afk19:44
*** kong has joined #openstack-tc20:28
smcginnisGlance was single vendor after ocata, briefly became multivendor in pike, now back again.20:48
*** ChanServ has quit IRC21:02
dmsimardsmcginnis: yeah, but it shouldn't be that way. Without Glance you can't do much, it's one of the core projects and should get (diverse) attention accordingly. I guess there's not much room for vendor implementation when compared to nova/neutron/cinder.21:07
*** ChanServ has joined #openstack-tc21:11
*** sets mode: +o ChanServ21:11
smcginnisdmsimard: I completely agree it should not be that way.21:12
smcginnisdmsimard: And I back that up even :)
* dmsimard nods21:13
cmurphyfrom that chart glance doesn't look that bad, keystone is about the same
mtreinishcmurphy: well except keystone has ~3x the reviews...21:24
persiaTo be fair, part of that is that glance is semi-stable at this point: while there are things that could go faster, there's not a lot of room for feature growth, etc. (in part because other projects depend on glance stability)21:34
smcginnisThat is true. There's some important work to be done there, but for the most part, glance is a mostly stable thing.21:36
mtreinishpersia: I disagree with that. It only feels semi-stable, it's not because of a conscious effort for stability, but more because of general inactivity21:36
persiamtreinish: That differs with the impression I've had over the past year or so, but my impression is largely second-hand.  I shall refresh my impression with the glance team when next I have the chance.21:38
persiaThat said, it may be that the limited number of folk working on the project has lead to a desire within the project for stability, to avoid overload, but some advertisement of what could be done might help with both of those things.21:38
mtreinishpersia: well part of it is the "feature" work the past few cycles is on things that most people don't actually use21:42
mtreinishand the generally low speed things move just makes it seem like nothing changes21:43
dmsimardEither that or refactors to improve things which aren't particularly exciting21:43
mtreinishdmsimard: it's not really that21:43
mtreinishheh, although I guess excitment level is a relative thing21:44
dmsimardmtreinish: I remember that being brought up last time glance contributions were discussed on the ML, I don't have first hand experience as I don't personally contribute to glance on a regular basis21:44
mtreinishdmsimard: well here's an example: the tasks api there is where a lot of work has been happening21:45
mtreinishwhich currently doesn't work with uwsgi as indicated in that backport patch21:45
mtreinishbut a lot of large operators don't care because it's not used (the only example I know of using it is rax)21:46
persiaI feel like this is a useful conversation, but that there are participants missing.21:49
cdentpersia: welcome to openstack?21:49
TheJuliacdent: Indeed21:50
dmsimardmtreinish: should I parse that as an expectation of operators to contribute back ?21:50
mtreinishdmsimard: they did, the backport was proposed by operators21:50
mtreinishand they're commenting on the patch saying they want the backport and indicating the tasks api isn't needed21:51
persiacdent: No.  While there is a lot of that in many fora, my very existence in open source largely depends on helping ensure conversations happen between the appropriate counterparties.  I believe that can happen with this conversation, although probably not until at least Thursday 14:00UTC, and possibly not until PTG.21:51
cdentpersia: I'm by no means suggesting that the conversation can't or shouldn't happen, simply that it is a common thing that the right people are often not in the right place at the right time. So I'm not sure what you mean by "No".21:52
persiaI also believe it should happen, as I have spent a fair amount of time interacting with the glance team over the past few years, and that there are features wanted that can be implemented and consumed by other projects is news to me (part of the no-new-feature effort has been related to v1->v2 transition, and worries about v3)21:52
persiacdent: By "No", I mean "I refuse to accept that one should have conversations without the appropriate counterparties because this is OpenStack.  I refute this as a part of our culture.  I shall rail against it, and do everything in my power to ensure that it is not accepted as commonplace."21:53
persiaMind you, I might fail :)21:53
cdentAh okay. I'm totally with you. My statement that it is common is very definitely _not_ an endorsement.21:54
persiaBe careful.  You carry a large reputation, and the dulcet tones of sardonicism do not always translate to IRC :)21:55
cdentI would have though my repucation would precede me on this particular topic; all I ever say is "let's have more useful conversations"21:56
smcginnisI wonder if our "top 5 help wanted" or "most wanted" or whatever it is now needs a little more, kind of like we started having champions for the goals.21:57
smcginnisIt seems like each item in the list should have a landing page, wiki or whatever, where interested people can go to get more details.21:57
smcginnisIt's one things to say "glance needs help" and leave it to the interpretor what that entails.21:57
smcginnisIt's another to have that go to a Glance Help Wanted page that lists specific areas where reviews are needed or where someone needs to do some design work for implementing something new that's needed or refactoring something existing.21:58
smcginnisWe need some guide rails to get these potential people more directly to the work we would like to see happen.21:58
cdentsmcginnis your comments raise a thought I've had tickling at the back of mind21:58
persiaMy fear with the top-5 is that it becomes a goal to be accepted in the 5.  It may be useful to redirect that energy into identifying what work could/should happen that can be attractive enough to acquire resources.21:59
cdentwe have a comment mental model which is that people rock up to openstack looking for something useful to do21:59
cdenthow many of the incoming contributors are of that form21:59
persiaAsking who-wants-what is a good start, rather than "this project (participant names rubbed off) needs help (for unspecified things)".21:59
cdentand how many are of the form "I was hired to work on openstack, some internally, some upstream"21:59
cdents/comment/common/ <- it's been a very long day of a lot of typing, more than usual21:59
smcginniscdent: I think it could help both types. And I think we do have both.22:00
cdentagree there are both22:00
cdentbut the distribution is something I don't understand22:00
persiacdent: Almost none.  The few of that form that I have met (who I tend to refer to as "CV polishers") usually seem to go away fast because the barrier to entry for getting something working and making a change is high (even ignoring getting the change accepted).22:00
smcginnisYeah, no insights there either.22:00
smcginnisI have a feeling we have more of the "I've been told to OpenStack" lately.22:00
persiaYes.  I think the common new contributor is either "I've been told to OpenStack" or "I've been told to ${project}".22:01
cdentin which case the help wanted list needs to be delivered to managers not contributors. that's something we've said before.22:02
persiaBut this goes to my long-standing dissatisfaction with contributing orgs: I do wish they would tell these people *why*, so that there was some concrete goals they could try to accomplish.22:02
persiamtreinish: Relaying messages: are you free Thursday @ 14:00UTC, or would you prefer an in-person discussion?22:03
mtreinishpersia: 1400utc is a bit early for me22:07
mtreinishbut I'll definitely be at the ptg22:07
persiamtreinish: No worries.  We'll have the discussion then.22:08
*** rosmaita has quit IRC23:07
*** kumarmn has quit IRC23:15
fungiyeah, i expect chinese new year has a lot of people otherwise occupied this week23:23
*** kumarmn has joined #openstack-tc23:38
*** cdent has quit IRC23:43
*** kumarmn has quit IRC23:44
*** kumarmn has joined #openstack-tc23:55
*** kumarmn has quit IRC23:57

Generated by 2.15.3 by Marius Gedminas - find it at!