14:02:04 <nikhil_k> #startmeeting Glance Artifacts sub-team meeting
14:02:04 <openstack> Meeting started Mon Jul  6 14:02:04 2015 UTC and is due to finish in 60 minutes.  The chair is nikhil_k. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:05 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:02:07 <openstack> The meeting name has been set to 'glance_artifacts_sub_team_meeting'
14:02:08 <ativelkov> o/
14:02:14 <mfedosin> o/
14:02:15 <dshakhray> o/
14:02:19 <nikhil_k> #chair ativelkov
14:02:20 <openstack> Current chairs: ativelkov nikhil_k
14:02:26 <nikhil_k> #topic agenda
14:02:26 <kragniz> o/
14:02:26 <sigmavirus24> o/
14:02:28 <nikhil_k> #link https://etherpad.openstack.org/p/glance-artifacts-sub-team-meeting-agenda
14:02:31 * kragniz lurks
14:02:58 <nikhil_k> ativelkov: I guess we have some updates for people to be aware about?
14:03:45 <nikhil_k> Besides that, I wanted to quickly discuss what do we mean by FastTrack and what we need to be carefuly about while merging patches.
14:04:01 <ativelkov> nikhil_k: not much. One of my artifact-related commits to oslo.versioned_objects broke nova and cinder tests last week, so I mostly was fixing that
14:04:27 <nikhil_k> gotcha. thanks.
14:04:28 <ativelkov> So, the first oslo.vo-related patch to glance will be submitted today
14:04:39 <ativelkov> I'll share the link when it's ready for reviews
14:04:40 <nikhil_k> cool
14:05:06 <nikhil_k> In that case..
14:05:09 <nikhil_k> #topic Define: FastTrack
14:05:54 <ativelkov> yup
14:06:33 <nikhil_k> To be carefuly about not breaking OS wide acceptance and ensuring we are compliant to the rest of the community standards, I believe we need to jot this down a bit
14:06:50 <nikhil_k> Somewhat along the lines of the infra manual
14:07:29 <nikhil_k> Basically, I find it really important that we include and respect changes in a particilar commit that are related to the following items (must NOT miss):
14:07:52 <nikhil_k> 1. Upgrades (document and clearly identity the right flags)
14:08:20 <nikhil_k> 2. Config changes (irrespective of the updates. documented and ensure that we keep them usable for operators)
14:08:42 <nikhil_k> 3. API changes (breaking or otherwise, if they impact API)
14:09:17 <nikhil_k> (above includes schema changes)
14:09:21 <ativelkov> nikhil_k: does p3 include the case if the change impacts only v3 part (i.e. experimantal API) -
14:09:26 <ativelkov> ?
14:09:46 <nikhil_k> 4. DB schema and migration changes
14:10:14 <nikhil_k> 5. notification, authN, authZ _standards_ changes in the project
14:11:01 <nikhil_k> ativelkov: yes please. We don't need big additions just those that indicate what and where things have changed
14:11:54 <ativelkov> So, we put ApiImpact flag anyway? Got it
14:11:58 <nikhil_k> ativelkov: this will also keep history fetch easily available and avoid commit reverts and extreneous conversations
14:12:25 <nikhil_k> We then know for what reasons which changes has been made
14:12:30 <nikhil_k> have*9
14:13:18 <sigmavirus24> Has there been any progress with the API-WG on the API spec for artifacts?
14:13:29 <sigmavirus24> (just a slightly tangential question, sorry)
14:13:31 <nikhil_k> Those are the ones I could think about.
14:13:52 <nikhil_k> sigmavirus24: yes. thanks for bringing that up.
14:14:12 <nikhil_k> I think we can just set a date to start the conversation to void getting delayed in Liberty
14:14:19 <mfedosin> just after the client I will start to work on api
14:14:51 <mfedosin> On Wednesday I suppose
14:15:02 <ativelkov> yup, I wanted to have this conversation already started, but oslo-related changes take too much time, unfortunately
14:15:15 <nikhil_k> To keep everyone aware, these conversations can happen for more than a week or two so, we can start small and start omre prep as we move along
14:15:42 <sigmavirus24> Also, if you can't make it to IRC meetings, please send discussion to the mailing list
14:15:49 <sigmavirus24> (for the API design that is)
14:16:22 <sigmavirus24> The reason I bring that up is because we've already started merging API code, including routes, and any updates that need to be made to conform to the spec will need an APIImpact flag I guess
14:16:44 <sigmavirus24> Which seems a little odd since v3 isn't really released, unless people are working against the liberty 1 milestone (8.0.0b1 for us)
14:17:47 <nikhil_k> 11.0.x ?
14:17:52 <sigmavirus24> er yeah
14:18:03 <sigmavirus24> sorry, dealing with keystone a lot lately has my version numbers confused
14:18:13 <nikhil_k> cool, np.
14:18:45 <nikhil_k> We have some push back on keeping Liberty v3 EXPERIMENTAL however, it would be nice to discuss that option with the WG
14:18:50 <ativelkov> I'll really try to get it started this week
14:18:53 <mfedosin> btw discussion mailing list turns into hell, because everyone has own opinion. Can it be more private?
14:19:35 <nikhil_k> mfedosin: we will have discussion in the open and get to the state that would make most sense
14:20:18 <mfedosin> be prepared for hell in the mailing list then :)
14:20:20 <nikhil_k> we can't keep everyone happy however, opinions can be accomodated in a rational perspective and not political interests
14:20:21 <ativelkov> The problem is that most of our "controversal" API design decisions were made bacause of other peoples' opinion :)
14:21:02 <sigmavirus24> ativelkov: there seemed to be a lot of consensus about the feedback left on the spec in the meeting(s) that I was in where it was discussed
14:21:07 <ativelkov> i.e. the separations into different resource collections for artifacts in different states were made because Mark has aked for it
14:21:10 <nikhil_k> ativelkov: yeah, that's one thing project leader(s) need to take the hit on
14:21:32 <sigmavirus24> i.e. you probably just listend to the wrong people =P
14:21:46 <nikhil_k> or people changed :-)
14:21:47 <ativelkov> sigmavirus24: :) Well, Mark was a PTL at that time ))
14:22:06 * sigmavirus24 is just teasing a bit
14:22:10 <nikhil_k> :)
14:22:22 <nikhil_k> understandable situation
14:22:35 <ativelkov> OK, I'll try to write a short summary of our API issues this week
14:22:45 <mfedosin> Well, we now have greater competence too
14:23:20 <ativelkov> When is API WG IRC meeting?
14:23:29 <mfedosin> rather then listen to other people advise we can give our own :)
14:23:51 <nikhil_k> #info API WG meeting sched http://eavesdrop.openstack.org/#API_Working_Group
14:24:20 <ativelkov> wow. 3 am my time this week
14:24:37 * mfedosin puts it in his calendar
14:24:38 <ativelkov> next week is much better
14:24:46 <nikhil_k> ativelkov: we can sync tomorrow and start it there
14:24:55 <sigmavirus24> ativelkov: yeah, we should either start discussing on the spec or the ML
14:24:56 <sigmavirus24> one or the other
14:24:57 <mfedosin> I don't sleep at this time anyway
14:24:59 <sigmavirus24> (or both)
14:25:02 <nikhil_k> ativelkov: usually people take their time and give more feedback in the following meeting
14:25:09 <ativelkov> got it
14:25:35 <nikhil_k> sigmavirus24: does that work? ^
14:26:07 <sigmavirus24> nikhil_k: it's already been 2 or three meetings since people reviewed the spec
14:26:20 <sigmavirus24> They'll probably re-review it when it makes sense to
14:26:42 <nikhil_k> yeah, but digesting our feedback adn coming up with rational usually takes time afaik
14:27:20 <nikhil_k> Best is getting references, that usually solves the purpose
14:27:30 <nikhil_k> #topic Glance client initial support
14:27:36 <mfedosin> folks, we have 3 minutes left. I want to talk about the client
14:27:56 <nikhil_k> mfedosin: by all means :-)
14:27:56 <mfedosin> It's almost done and there are several commits
14:28:10 <nikhil_k> #link https://review.openstack.org/#/c/189860/
14:28:36 <mfedosin> I need one day to put it in a order
14:28:52 <mfedosin> who can I ping to review them?
14:29:03 <nikhil_k> o/
14:29:18 <mfedosin> tomorrow or on Wednesday
14:29:22 <nikhil_k> np
14:29:28 <mfedosin> cool, thank you
14:29:39 <mfedosin> we need the client to begin writing tests
14:29:49 <nikhil_k> tempest?
14:30:18 <ativelkov> I'd prefer to start bothering with tempest only when the API is STABLE
14:30:19 <mfedosin> ostf I think
14:30:31 <nikhil_k> agree with ativelkov
14:30:36 <mfedosin> and tempest next
14:30:39 <nikhil_k> it can make things ugly
14:30:41 <ativelkov> or at least approved by API WG
14:30:45 <nikhil_k> yep
14:30:50 <nikhil_k> ok, we are out of time
14:31:05 <mfedosin> thanks folks
14:31:06 <nikhil_k> Let's discuss in the prj channel if needed
14:31:08 <nikhil_k> thanks!
14:31:11 <nikhil_k> #endmeeting