19:00:19 <catherineD> #startmeeting refstack 19:00:19 <openstack> Meeting started Tue Aug 23 19:00:19 2016 UTC and is due to finish in 60 minutes. The chair is catherineD. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:00:23 <openstack> The meeting name has been set to 'refstack' 19:01:18 <pvaneck> o/ 19:02:45 <hogepodge> o/ 19:03:15 <catherineD> #link meeting agenda and notes, https://etherpad.openstack.org/p/refstack-meeting-16-08-23 19:05:19 <catherineD> Let's start 19:05:51 <catherineD> seems like we will have a short meeting today 19:05:58 <catherineD> #topic Implement additional properties Defcore waiver 19:06:40 <catherineD> #link Implement additional properties Defcore waiver https://review.openstack.org/#/c/343022/ 19:07:52 <pvaneck> #link https://review.openstack.org/#/c/349213/ 19:07:57 <hogepodge> is luzC around? 19:07:59 <pvaneck> that was the refstack spec that was abandoned 19:08:02 <catherineD> For this one hogepodge: 's review is most important, since hogepodge: knows the use well 19:08:11 <catherineD> pvaneck: thx 19:08:24 <luzC> hey 19:08:28 <luzC> yes... 19:08:39 <hogepodge> I just got back from vacation, so I haven't had a chance to review yet. 19:09:53 <catherineD> hogepodge: np 19:10:11 <luzC> @hogepodge ok, also if you have 30 min this afternoon we can go through the other options so that you can give me your feedback and agree on a course of action 19:10:40 <luzC> either as it is on the spec or other alternative option 19:11:12 <hogepodge> luzC: Definitely do, I have the board meeting following this, but am free after 3 PT 19:11:13 <catherineD> luzC: I added couple comments .. but I think hogepodge: knows the use cases well .. so hogepodge: 's input is the most importnat ... we can review next week 19:11:37 <hogepodge> at first glance, a bunch of whitespace issues, but the central problem seems to be addressed 19:11:52 <catherineD> pvaneck: thx for the correct link .. I will update the agenda 19:12:02 <luzC> sounds good thank you @hogepodge @catherineD 19:12:56 <catherineD> moving on 19:13:10 <catherineD> #topic Procduct version specification 19:13:43 <catherineD> #link Specitication to support product model with multi-versions. ( https://review.openstack.org/#/c/353903/ ) 19:15:32 <catherineD> Our goal is to be able to associate results data to vendor products in the Newton cycle as it is shown on the prototype 19:16:03 <pvaneck> yea, this model was used in the prototype, and I had no issues with it 19:16:04 <catherineD> Without this spec merge, all development work are on hold ... 19:16:41 <catherineD> we have the code and prototype ready since August 1st, 2016 ... 19:16:58 <catherineD> but we have not made any progress since 19:17:27 <pvaneck> just need to address the issues inline, then i will +2 19:18:11 <catherineD> pvaneck: thx 19:18:25 <hogepodge> I don't see any -1, so it seems like moving forward is ok. 19:19:39 <catherineD> I will push a new patch address pvaneck: 's comments ... 19:19:55 <pvaneck> and chris's comments too 19:20:19 <catherineD> pvaneck: hogepodge: sure 19:20:30 <catherineD> after that could you all review again 19:20:43 <pvaneck> yep 19:21:31 <hogepodge> will do 19:21:43 <catherineD> thank you! 19:22:53 <catherineD> I have specifically asked Andrey and Sergey to attend this meeting so that we can make decision ... I will connect with them afterward 19:23:25 <catherineD> moving on .. 19:23:37 <catherineD> #topic Pending reviews 19:24:26 <catherineD> we have a few pending reviews .. However, all of those will be affected depending on the outcomes of https://review.openstack.org/#/q/openstack/refstack,n,z 19:24:54 <catherineD> so I think we should postpone reviewing of those to later 19:25:54 <pvaneck> sure 19:26:15 <catherineD> anything else to discuss? 19:26:56 <pvaneck> nothing from me 19:26:59 <catherineD> if not, we have a short meeting today 19:27:07 <andrey-mp> o/ 19:27:17 <catherineD> andrey-mp: Hi 19:27:18 <andrey-mp> hi all 19:27:53 <catherineD> andrey-mp: are you able to see the back log 19:28:08 <catherineD> we are about done .. 19:28:17 <andrey-mp> ah, ok ) 19:28:19 <catherineD> #link meeting agenda and notes, https://etherpad.openstack.org/p/refstack-meeting-16-08-23 19:29:08 <catherineD> andrey-mp: please review https://review.openstack.org/#/c/353903/ 19:29:18 <andrey-mp> I saw it. 19:29:52 <andrey-mp> i posted one comment. and now I'm waiting for answers/updates due to comment from other reviewers 19:30:29 <catherineD> I think I address all comments ... andrey-mp: do you mind remind me which one is that? 19:31:59 <andrey-mp> oh sorry, i didn't post it (just wrote) https://review.openstack.org/#/c/353903/5/specs/newton/approved/product-version-datamodel-api.rst 19:32:10 <catherineD> andrey-mp: :-) 19:32:16 <andrey-mp> my last comment was very small 19:32:31 <andrey-mp> and new comments are more interested 19:33:10 <catherineD> andrey-mp: I saw it ... let's ask hogepodge: 19:34:34 <hogepodge> We don't want a situation where you can create confusion by grabbing another vendor's name, but it's not necessarily critical for this review 19:34:36 <catherineD> hogepodge: are product name unique per vendor a requirment from OSF? If not I can remove that sentence .. 19:35:01 <catherineD> hogepodge: I think we type at the same time .. 19:35:04 <hogepodge> If a malicious or even ill-informed agent were to create a name collision, I think we could revisit the issue 19:35:48 <catherineD> hogepodge: andrey-mp: OK in that case, I can remove the sentence and make a comment there 19:35:56 <hogepodge> I think we see things like VendorA OpenStack, an VendorB OpenStack. If they just use "OpenStack" as the name we wouldn't want to block everyone else 19:36:45 <andrey-mp> my comment was only about this line in this review 19:37:02 <catherineD> hogepodge: Should VendorA has more than one prodcut named "OpenStack" ? 19:37:24 <andrey-mp> this review is about product-version relation and is not about constraints in vendor-product tables 19:37:38 <hogepodge> catherineD: in practice it's the smaller vendors who do something like that, but it's the source of name collisions. 19:37:59 <hogepodge> to answer andrey-mp's concerns on it, I don't object to removing the line and worrying about it elsewhere 19:38:32 <catherineD> andrey-mp: if OSF have this requirement then we will implement constraint 19:38:55 <catherineD> OK I will remove the line 19:39:22 <catherineD> I will push a new patch after this meeting ... 19:39:46 <catherineD> please review 19:40:02 <catherineD> anything else to discuss? 19:40:20 <andrey-mp> catherineD: this is old discussion about requirements ) it's better to have all requirements before architect system and develop it (and it's not related to agile or other methodlogies) 19:41:21 <andrey-mp> but right now we have a situation when new requirements are born every month :) 19:41:27 <catherineD> andrey-mp: I understand ... however giving the short runway we have to have data associating to vendor product ... I don't it is realistic 19:42:08 <hogepodge> andrey-mp: I think you are mischaracterizing the situation, and the requirements for defcore have been expressed multiple time, including in the last meeting at your request. again 19:42:43 <hogepodge> andrey-mp: as part of the project mission of supporting defcore, we've met at summits and mid-cycles to discuss these plans and move the project forward 19:43:04 <catherineD> I check on other projects .. most are done incrementally with a over target such as for us associating data to vendor 19:43:17 <andrey-mp> hogepodge: this mischaracterizing was born because i can't find a one official place with all requirements 19:43:49 <catherineD> andrey-mp: I don't think you can find one official place with all requirements 19:43:59 <catherineD> in most of OpenStack project 19:44:06 <catherineD> you will find many specs 19:44:07 <andrey-mp> no one can collect all of them from IRC logs, review, specs, ... 19:44:41 <hogepodge> andrey-mp: and yet you ask, and we describe, and write specs and plan, and you're saying we're not designing properly. We had a huge discussion about this, and documented it, at last defcore mid-cycle 19:44:50 <catherineD> that is why we need to break up the requirements and document them in the spec just like other project did 19:45:23 <andrey-mp> hogepodge: could you get me a link to this document? 19:45:29 <hogepodge> andrey-mp: and you have not been an active participant in any defcore efforts, except for one review. You don't show up at meetings, and you don't show up to midcycle, yet you want to work on a project where the mission is to support defcore efforts 19:46:07 <hogepodge> the google doc captured the data model, the spec catherine posted, the last meeting minutes 19:47:02 <catherineD> andrey-mp: we tried to document all discussion in the working session to specs 19:47:06 <andrey-mp> yeah, some time ago I understood that refstack is just a project to support defcore efforts 19:47:18 <sslypushenko> o/ 19:47:59 <hogepodge> andrey-mp: This project is to support defcore http://git.openstack.org/cgit/openstack/governance/tree/reference/projects.yaml#n5624 19:48:25 <catherineD> andrey-mp: It is still a project to support DefCore efforts ..that is its focus ... but if we can at the same time help other projects that would be great 19:48:36 <andrey-mp> but hoped that this is a 'RefStack is a source of tools for OpenStack interoperability testing.' 19:48:56 <andrey-mp> as posted on the main page of RefStack 19:49:27 <catherineD> andrey-mp: we all want RefStack to grow to that but we need to support DefCore first 19:49:59 <catherineD> andrey-mp: supporting DefCore is supporting :OpenStack interop testing 19:50:21 <andrey-mp> catherineD: it's sad that we didn't catch this early 19:50:55 <catherineD> andrey-mp: why is that? 19:51:51 <andrey-mp> Alex tried to do something to improve interoperability testing 19:51:53 <catherineD> andrey-mp: I think the question should be asking is ... if RefStack supports DefCore (this is a giving) and at the same time support others . Isn't that a win-win situation? 19:51:54 <hogepodge> andrey-mp: and we are taking active steps to support general interoperability across openstack, but again, except for a major version patch dropped on defcore (which we're reworking to meet defcore working group goals), you have not participated in a meaningful way. We are working towards general interoperability in the second half of this year, but refstack right now needs to support finer-grained vend 19:51:54 <hogepodge> or details for test result reporting 19:52:41 <catherineD> andrey-mp: you have not prove to me that what we are doing for DefCore is not applicable to your goals 19:52:55 <catherineD> or break you 19:53:30 <andrey-mp> please stop attack me :) 19:53:33 <hogepodge> andrey-mp: I'm sorry if the defcore timeline doesn't match up with what you need, but I met personally with Randy to discuss these plans and he seemed ok with the path we were taking. I expect to meet the needs of more general interop testing for OpenStack projects late this year and early next, all within the defcore framework and with board approval 19:54:51 <andrey-mp> I'm not participating in DefCore meetings because it's not my role 19:55:31 <andrey-mp> I'm just trying to implement some that was discussed in Austin 19:55:44 <catherineD> andrey-mp: Please do no think that about attack ... we are a team ... I would like RefStack to support interop tests (that is DefCore and others ) that is why we keep the discussion going 19:56:32 <catherineD> andrey-mp: what we discuss in Austin is still relevent ... it is the detail that was not disucssed in Austin (in this case it is product has versions) 19:57:33 <catherineD> the domain model is still intact ... it is the product definition that needed to be improved ... 19:59:04 <catherineD> this is why I need to understand how implement product with version the way I defined would break the model discussed in Austin .. .because in my mind it enhance the model 19:59:06 <andrey-mp> catherineD: I thought that domain model suggests that each version is an new product. And I thought we discussed it Austing in two words 19:59:28 <catherineD> andrey-mp: domoain model has no mention of version 20:00:10 <hogepodge> I can not stay beyond the meeting end time. 20:00:13 <catherineD> anyway we are out of time .. 20:00:20 <catherineD> Thank you all ... 20:00:52 <catherineD> I will push a new patch please review .. we need to move the project forward 20:01:08 <catherineD> #endmeeting