18:59:17 <davidlenwell> #startmeeting refstack 18:59:18 <openstack> Meeting started Mon Feb 23 18:59:17 2015 UTC and is due to finish in 60 minutes. The chair is davidlenwell. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:59:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:59:22 <openstack> The meeting name has been set to 'refstack' 18:59:29 <hogepodge> o/ 19:00:28 <davidlenwell> agenda: recap from face to face, uuid progress, pending reviews, news from david, open discussion 19:00:58 <pvaneck> o/ 19:02:02 <vladiskuz_> o/ 19:02:59 <catherineD> o/ 19:03:52 <zehicle> o/ 19:04:02 <davidlenwell> okay .. I think we have enough to get started.. 19:04:14 <davidlenwell> #topic recap from face to face 19:04:34 <catherineD> https://etherpad.openstack.org/p/refstack_f2f_feb_2015 19:04:43 <davidlenwell> aww you beet me to it ;) 19:04:57 <davidlenwell> We covered a lot of ground on wednesday.. 19:05:09 <zehicle> #link https://etherpad.openstack.org/p/refstack_f2f_feb_2015 19:05:30 <sslypushenko__> o/ 19:05:40 <davidlenwell> We started with reviewing the stated refstack usecases that zehicle and I came up with together when we kicked this thing off 19:06:02 <davidlenwell> decided what was still relevant and what was already addressed 19:06:37 <davidlenwell> Then we took a hard look at what was still remaining 19:07:25 <davidlenwell> we then tried to decide on our next set of goals.. 19:08:04 <davidlenwell> if you scroll towards the bottom you can see what we decided the next set of priorities are.. each item lists the names of folks of people who are likely to do the work.. 19:08:14 <davidlenwell> we tried to make sure everyone has stuff to do .. 19:08:51 <davidlenwell> Please take the time to review this doc if you haven't already and feild any questions to catherine or myself in channel 19:09:53 <zehicle> davidlenwell, do we have this prioritized? 19:10:03 <zehicle> is it roughly in order? 19:10:33 <davidlenwell> roughly it is in order yes 19:10:54 <zehicle> ok 19:10:57 <davidlenwell> however none of it is set in stone 19:11:16 <zehicle> understood 19:11:21 <zehicle> good to have use c ases :) 19:11:22 <catherineD> zehicle: and we try to have all of us having work items in each priorrity item 19:11:28 <sslypushenko__> What doc we are talking about? 19:11:44 <hogepodge> #link https://etherpad.openstack.org/p/refstack_f2f_feb_2015 19:12:30 <davidlenwell> I think thats enough time on that.. I think everyone is up to speed and if anyone wants to discuss it . we can do it in channel .. 19:12:45 <davidlenwell> #topic uuid in tempest status.. 19:12:59 <sslypushenko__> I have a good news) 19:13:06 <davidlenwell> hogepodge: would you like to update us? 19:13:06 <hogepodge> https://review.openstack.org/#/c/157273 19:13:18 <sslypushenko__> It looks like tagging script woks) 19:14:00 <hogepodge> That's good news. 19:14:12 <davidlenwell> excelent! 19:14:16 <hogepodge> Pretty much all of the work is done. Just need to tag the tests and submit for review. 19:14:19 <sslypushenko__> I have finished it a couple minutes ago 19:14:32 <hogepodge> sslypushenko__: did you catch the pep8 failures? 19:14:34 <sslypushenko__> I did minimal set of manual tests 19:14:46 <hogepodge> (unrelated to uuid) 19:15:01 * zehicle excited to see the UUID progress! 19:15:04 <sslypushenko__> I didn't fix pep8 issues 19:15:06 <hogepodge> also, did you see the comment on running the script? 19:15:13 <sslypushenko__> I will do it after meeting 19:15:42 <hogepodge> So it sounds like mostly just tidying up then reviewing. I send an e-mail to openstackdev to alert everyone about the big push and changes is means for future tests. 19:16:15 <hogepodge> (that sentence barely makes sense, but you get the idea) :-P 19:16:56 <hogepodge> So feel free to review and so on. Thanks sslypushenko__ 19:16:59 <davidlenwell> well thats aweomse news! 19:17:07 <davidlenwell> *awesome 19:17:11 <sslypushenko__> hogepodge I saw your review. We can disscuss detail later in refstack channel 19:17:34 <Rockyg> o/ 19:19:23 <hogepodge> That's about it for that topic I think. 19:19:56 <davidlenwell> I am very happy with this progress.. Very good job! 19:20:06 <davidlenwell> lets move on and go over pending reviews 19:20:11 <davidlenwell> #topic pending reviews 19:20:55 <davidlenwell> #link https://review.openstack.org/#/c/153734/ 19:21:58 <davidlenwell> lets get eyes on this one.. its been pending for a while 19:22:59 <sslypushenko__> I was pretty focused on uuid script. I will do review as soon as possible 19:23:08 <davidlenwell> thanks sslypushenko__ 19:23:08 <catherineD> I am testing ... 19:23:16 <davidlenwell> while you are at it .. review this #link https://review.openstack.org/#/c/157144/ 19:23:25 <hogepodge> I'll look too. 19:23:54 <davidlenwell> hogepodge: that last one I linked should be of interest to you.. its what we talked about in the january face to face 19:24:01 <hogepodge> On the second it's going to get complicated for tests that have moved. 19:24:22 <catherineD> vladiskuz_: did a good job in reviewing https://review.openstack.org/#/c/157144/ 19:24:48 <catherineD> hogepodge: why is that? 19:25:20 <Rockyg> one quick comment from defcore f2f: we are moving away fro using the word "core" zehicle, what was the new word? 19:25:45 <davidlenwell> Rockyg: thats a bit off tpoic .. lets save it for open discussion 19:26:01 <hogepodge> catherineD: If you're listing the tests and searching, but the name has changed, you can miss it. So mostly from the pov of leaving gaps 19:26:09 <Rockyg> will need to update what we use instead of --core in the last review 19:26:09 <davidlenwell> because you'll have to scan for the test if all you have is the uuid.. we don't know its path so launching it isn't as easy 19:26:22 <vladiskuz_> catherineD: Thanks! But I think this patch need more clarification 19:26:24 <catherineD> hogepodge: that is why we are now settle on tag-3 19:26:24 <davidlenwell> Rockyg: please review the code and mention that 19:26:55 <hogepodge> catherineD: What are we going to do about failing neutron tests? We have to work from head in those cases. 19:27:10 <hogepodge> (that is tests that fail because of neutron) 19:27:11 <catherineD> hogepodge: what are failling neutron test? 19:27:30 <hogepodge> tests that fail with the "no default network specified" bug 19:27:48 <zehicle> I was going to do updates all together 19:28:01 <catherineD> hogepodge: I am not aware of that .. Let's discuss later 19:28:10 <zehicle> Rockyg, required capabilities 19:28:19 <hogepodge> catherineD: It's not a show stopper, though. There are workarounds. 19:28:26 <Rockyg> thanks 19:29:10 <davidlenwell> okay .. we have some other pending reviews from vladiskuz_ mostly around test deployment and database migrations ... lets get eyes on them 19:29:49 <vladiskuz_> This https://review.openstack.org/#/c/156569/ and this https://review.openstack.org/#/c/156983/ 19:30:08 <davidlenwell> yes .. thanks for linking vladiskuz_ 19:30:25 <catherineD> vladiskuz_: davidlenwell: will review those today .. 19:30:27 <vladiskuz_> Can we add new gate in review for functional tests? 19:30:37 <davidlenwell> sure 19:30:41 <catherineD> back to https://review.openstack.org/#/c/157144/ 19:30:56 <davidlenwell> vladiskuz_: do you want to do that commit? 19:30:59 <catherineD> let have discussion on #refstack about https://review.openstack.org/#/c/157144/ 19:32:17 <hogepodge> I like that patch, especially if it allows for defining custom capabilities files. 19:32:20 <catherineD> because https://review.openstack.org/#/c/157144/ cover one of the use case (1.4 in the f2f) 19:33:31 <vladiskuz_> For this patch https://review.openstack.org/#/c/156569/ will be good if we can run tests in our review system with tox - e func 19:33:57 <Rockyg> vladsjuz_ ++ 19:34:50 <vladiskuz_> and add gate, for example 'gate-refstack-mysql-functional' or something like this 19:35:28 <davidlenwell> lets make it non-voting but yes .. I think its a good gate to add 19:36:40 <vladiskuz_> davidlenwell: why non-voting? 19:37:17 <Rockyg> vladiskuz_: at least until it shows it's solid 19:37:25 <davidlenwell> that ^^ 19:38:01 <davidlenwell> lets move on for now.. we can circle back to this later 19:38:46 <davidlenwell> #topic news from David 19:39:07 <davidlenwell> It wasn't an easy choice to make.. But I have resigned my position at blue box.. in the past months as many of you know it has been hard to strike a balance within the company between upstream development and internal product work.. at the end of the day I had to prioritize community work and leave blue box .. No word yet on where I might go.. but I am committed to staying in OpenStack and working upstream and I remain committed to 19:39:07 <davidlenwell> Refstack. 19:40:36 <zehicle> ok, thanks for the update 19:41:13 * zehicle keeps an eye out for davidlenwell positions 19:41:26 <davidlenwell> The good news from this is that I will no longer be distracted by internal product work and will be able to put in more hours on refstack. 19:41:30 <davidlenwell> thanks zehicle 19:41:52 <Rockyg> Yay! more progress! 19:42:45 <davidlenwell> basically I've decided not to take any job until after the summit.. I want to put all my energy into meeting the goals we layed out in our face to face last week 19:42:46 <Rockyg> Uh, put together what you want in a position -- area(s) you want to work in -- and let us know. Nothing here (it's a mess) but, who knows what we may hear. 19:43:41 <davidlenwell> okay . we can more on now to open discussion 19:43:51 <davidlenwell> #topic open discussion 19:43:51 <zehicle> I can give an update fro DefCore 19:44:03 <davidlenwell> yes . lets do that 19:44:04 <zehicle> mostly, we talked about DefCore process stuff 19:44:14 <zehicle> but there's a potential change that would make Refstack easier 19:44:41 <zehicle> basically, we are suggesting a move away from Capabilities being RELEASE based to being time based 19:44:42 <davidlenwell> oh? 19:45:02 <davidlenwell> thats very interesting 19:45:10 <zehicle> so there's ONE capabilities score JSON that gets updated incrementally based on the process 19:45:13 <catherineD> zehicle: what does that mean? 19:45:22 <zehicle> you'd pick the date of the file that you apply 19:45:31 <zehicle> and Vendors would have to be within the 2 most recent 19:45:54 <Rockyg> Yeah. It made sense. You post which test set(s) you're compliant with 19:46:00 <hogepodge> So rather than say "icehouse capabilities, juno capabilities, etc..." 19:46:04 <zehicle> so, if the Board approves DefCore 2015.May, then that would be the tests you check against 19:46:10 <zehicle> it really works w/ the UUID concept 19:46:12 <davidlenwell> I like it 19:46:43 <hogepodge> It also matches with the idea of "versionless" tempest 19:46:43 <zehicle> we talked about it for a long time. In the end, we felt that it felt more like a "OpenStack Spec" than the previous naming 19:46:49 <sslypushenko__> Sounds interesting 19:46:59 <davidlenwell> we'll have to have some discussion on impact 19:47:19 <zehicle> it also helped explain that the DefCore list is trailing. So you don't expect to see capabilities from the latest release 19:47:51 <zehicle> for example, the DefCore 2015.7 guideline would have no Kilo capabilities 19:47:55 <Rockyg> and I don't think it will affect the current tags we are looking at, just might require more? 19:48:00 <zehicle> (unless it was something added to catch up) 19:48:01 <hogepodge> Also, something kind of related 19:48:38 <zehicle> oh, and Rockyg, you were trying to remember "Guideline" 19:48:43 <zehicle> we did not want to call it a spec 19:49:10 <hogepodge> I put forward an idea to help identify capabilities, and going to try and do a pilot with it on Keystone. 19:49:47 <zehicle> does that make sense? You all are our first test run of explaining the new approach 19:50:03 <Rockyg> there's also now a timeline for defcore to get advisory and final capabilities and tests approved by board. 19:50:32 <davidlenwell> I think I get it .. I like the implications .. 19:50:33 <zehicle> yes. very accelerated one 19:50:56 <zehicle> we are trying to have guidelines published in March, April and May to catch up 19:51:07 <zehicle> Then do to a 6 month schedule 19:51:31 <Rockyg> with advisory in the middle of the cycle 19:51:44 <zehicle> we had great participation from the Foundation staff at the F2F - they were helpful in getting this lined up 19:51:52 <Rockyg> very 19:51:53 <zehicle> in addition to hogepodge, of course. 19:52:02 <hogepodge> (heh) 19:52:06 <catherineD> so what happen to beta tests? 19:52:08 <Rockyg> He's foundation staff 19:52:12 * zehicle hugs hogepodge 19:52:19 * hogepodge blushes 19:52:43 <zehicle> there's a concept of required, depricated & advisory capabiltities 19:52:44 <Rockyg> so, advisory is a proposed list that allows time to shake out issues for final 19:52:56 <zehicle> and we also can flag tests 19:52:58 <zehicle> Rockyg, +1 19:53:21 <Rockyg> advisory capabilities list (comes out 3mos before final) 19:53:34 <hogepodge> At the next qa meeting I'm going to make a proposal for a new type of test tag 19:53:39 <catherineD> The earlier plan is to have beta tests with some vendors in March ... is that still in plan? 19:53:44 <hogepodge> To help out with this in the long run too. 19:53:46 <zehicle> Rockyg, I think we were going to use "draft" 19:53:59 <zehicle> and then use advisory inside the doc 19:54:01 <Rockyg> Was that what we ended up with? OK. 19:54:10 * zehicle thinks draft sounds more spec like 19:54:15 <Rockyg> The beta is still very important 19:54:36 <hogepodge> [interop] tag that has a requirement of being an api test that does not require admin access, and uses only a minimal set of configuration (credentials, endpoint, image-id, network-id, for example) 19:54:39 <Rockyg> We need to know if the draft/final will work The beta will demo it 19:54:53 <zehicle> :) 19:55:01 <hogepodge> Going to try a pilot on keystone tests. (sorry to interleave, but running out of time) 19:55:38 <hogepodge> So we can start pulling capabilities for defcore from [interop] tagged tests. Longer term goal, but gets to the heart of cross-cloud testing by end users. 19:56:46 <hogepodge> Huge community interest in testing. Lots of proposed talks for the summit. 19:57:07 <Rockyg> It was also pointed out that there is still a need for admin at some level -- for test setup. We should Identify exactly what and document 19:57:09 <hogepodge> Go vote! https://www.openstack.org/vote-vancouver/presentation/testing-your-cloud-how-to-configure-your-cloud-and-tempest 19:57:30 <Rockyg> accounts and stuff 19:57:32 <hogepodge> Rockyg: If you need admin you can't have interop tag. 19:57:40 <catherineD> hogepodge: we shoudl send that line to fits 19:57:43 <hogepodge> But there can be interop nq 19:57:55 <hogepodge> (non qualifying) 19:58:25 <hogepodge> It should be set up before tests are run if it needs admin. 19:58:33 <hogepodge> (which is reasonable, cloud prep) 19:58:49 <davidlenwell> okay folks we are running out of time.. and another team uses this channel in two minutes .. lets wrap it up and continue discussion in #refstack 19:58:54 <Rockyg> Right. hogepdge. We need to document that *very* well 19:59:03 <Rockyg> k 19:59:06 <hogepodge> Thanks everyone! 19:59:11 <zehicle> good progress! 19:59:30 <davidlenwell> Thank you all for being here today .. attendance has been a lot better in the last month! 19:59:48 <davidlenwell> #endmeeting