19:00:40 <catherineD> #startmeeting refstack 19:00:41 <openstack> Meeting started Mon Apr 6 19:00:40 2015 UTC and is due to finish in 60 minutes. The chair is catherineD. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:42 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:00:44 <openstack> The meeting name has been set to 'refstack' 19:00:48 <hogepodge> o/ 19:00:49 <davidlenwell> o/ 19:00:50 <catherineD> roll call 19:00:53 <pvaneck> o/ 19:01:31 <catherineD> Today's agenda: 1) recap goals for Vancouver 2) design session planning 3) plan for upgrading refstack-client to latest tempest 4) pending review 5) open discussion 19:02:03 <hockeynut> o/ 19:02:23 <catherineD> #topic recap goals for Vancouver 19:02:49 <vladiskuz|2> o/ 19:03:11 <catherineD> Please look at the action items at the bottom in https://etherpad.openstack.org/p/refstack_f2f_feb_2015 19:04:12 <catherineD> our goal is being able to demo refstack.net (or refstacl.org) server with test content ... 19:04:15 <sslypushenko__> o/ 19:04:18 <Rockyg> o/ 19:04:42 <sslypushenko__> catherineD: +! 19:04:47 <catherineD> vladiskuz|2: 's API was merged last week .. 19:05:27 <vladiskuz|2> catherineD: thx! it's good :) 19:05:38 <catherineD> thx vladiskuz|2: now pvaneck: is workign on the presentation layer ... pvaneck: any update? 19:06:37 <pvaneck> yes, I have created the scaffolding for the site in angular 19:06:58 <hogepodge> John Oliver interviewed Edward Snowden. Amazing. 19:07:01 <sslypushenko__> pvaneck Sounds greate! 19:07:03 <pvaneck> working on getting it hosted on github pages for now 19:07:09 <davidlenwell> hogepodge: yeah .. that was very entertaining 19:07:11 <hogepodge> ugh, wrong room. Sorry. :-( 19:07:11 <pvaneck> in person? 19:07:24 <davidlenwell> yes .. in person 19:07:26 <hogepodge> sincere apologies 19:07:28 <pvaneck> haha 19:08:09 <catherineD> pvanpvaneck: thx ... Let us know when it is up so we can review .. 19:08:20 <vladiskuz|2> pvaneck: can you share you results with us? because I working on profile and auth with angular 19:08:56 <catherineD> sslypushenko__: vladiskuz|2: are working on the signed API to upload data ..sslypushenko__: has a patch that I will test today ... 19:09:50 <pvaneck> http://pvaneck.github.io is what I have for now, was currently testing out displaying capabilities using angular 19:09:53 <sslypushenko__> catherineD: I think upload is almost finished... Need a little polish on client side 19:10:01 <pvaneck> still needs work 19:10:16 <pvaneck> and content i guess 19:10:54 <sslypushenko__> pvaneck: It is really good start! 19:11:07 <pvaneck> would like if the refstack.net could be updated with the lasted api 19:11:07 <vladiskuz|2> pvaneck: looks good! 19:11:13 <pvaneck> latest* 19:11:18 <catherineD> pvaneck: Let us know when you are ready ... we can have a phone call so you can share what you have learned to jump start us ... so our team does not need to learn from scratch .. 19:11:52 <catherineD> I think all status looks good ... 19:12:04 <catherineD> we need to talk about where to host the refstack server ... 19:12:22 <catherineD> right now it is hosted privately by David ... 19:12:29 <davidlenwell> I'd like to follow up with clee and see if we can get some donated isntances from dreamhost 19:12:48 <Rockyg> I think we need to get openstack-infra involved. They need to be the ultimate hosts. 19:13:09 <davidlenwell> Rockyg: I am not sure that will be an option for at least another 6 months 19:13:10 <sslypushenko__> Rockyg +1 19:13:20 <davidlenwell> we need to host it some place neutral until then 19:13:34 <catherineD> davidlenwell: thx ... Let's us revisit this topic next week ... 19:14:04 <davidlenwell> sounds good 19:14:19 <catherineD> Rockyg: +1 ... I think vendor would like it to host on OpenStack eventually ... 19:14:38 <davidlenwell> sure .. thats always been the ultimate goal 19:14:39 <catherineD> any other thoughts on this topic? 19:15:17 <catherineD> #agreed revisit location to host refstack server in the next IRC meeting 19:15:45 <catherineD> alright 19:15:47 <davidlenwell> sounds good 19:15:54 <catherineD> #topic design session planning 19:16:15 <davidlenwell> we are supposed to be putting thoughts here.. https://etherpad.openstack.org/p/refstack-qa-vancouver-design-session 19:16:45 <davidlenwell> however as many of you know.. I've been on vacation since last week .. so I have not had the time to leave my thoughts here 19:16:58 <catherineD> davidlenwell: ya :-( 19:18:27 <catherineD> from IBM , we do not have our aprroved list to summit yet ... so we do not know ... 19:18:52 <catherineD> I know davidlenwell: hogepodge: will be at the summit ... who else? 19:19:13 <davidlenwell> I'm sure rob and Rockyg will be there .. 19:19:57 <catherineD> sslypushenko__: how about you? 19:20:15 <sslypushenko__> No, I'm not planning 19:20:44 <catherineD> How about Vladiskuz? 19:20:46 <Rockyg> yup. I'll be there. 19:21:03 <vladiskuz|3> catherineD: I'm not planning too 19:21:12 <catherineD> oh 19:21:44 <catherineD> I guess we will let people to update https://etherpad.openstack.org/p/refstack-qa-vancouver-design-session .. 19:22:03 <sslypushenko__> Too far away 19:22:24 <catherineD> about a month .. 19:22:59 <sslypushenko__> I meant in miles) 19:23:14 <hogepodge> I'd like for us to have a session that will make it possible for the folks who can't attend to somehow conference in. 19:24:04 <catherineD> hogepodge: +1 19:24:36 <sslypushenko__> It will be great 19:25:18 <catherineD> #agreed refstack team members will update https://etherpad.openstack.org/p/refstack-qa-vancouver-design-session for design session idea/topic 19:25:40 <catherineD> #topic plan for upgrading refstack-client to latest tempest 19:26:12 <catherineD> refstack-client is now at tag-3 ... we need to plan for an update to the latest tempest ... 19:26:24 <catherineD> question: do we wait for tag-4? 19:26:59 <davidlenwell> I think we should 19:27:42 <catherineD> I think tag-4 will be one week after Vancouver ... 19:28:02 <catherineD> any other opinion? 19:28:14 <Rockyg> Don't we need to support both? 19:28:26 <sslypushenko__> Latest? May be we should pick some commit which is close to date of latest capabilities approval? 19:28:29 <mtreinish> catherineD: I'll push it soon after the Kilo release 19:28:33 <davidlenwell> with uuid's on tests drifting between versions won't be as painful 19:28:52 <catherineD> mtreinish: thx you very much! That would help! 19:29:00 <sslypushenko__> davidlenwell: Unfortunately - it will 19:29:37 <sslypushenko__> Current implementation of uuids id not fits perfectly to refstack needs 19:29:52 <davidlenwell> sslypushenko__: can you elaborate ? 19:30:17 <mtreinish> catherineD: I sometimes wait for a last min patch to fix something before pushing the tag, but only if it's queued up to land soon after the release. So I expect it would be within a week of the kilo release 19:30:18 <catherineD> Rockyg: with branchless, whatever future tag should be backward competible...so we only need to test with one tag release 19:31:10 <davidlenwell> I think waiting until the new tag is the way to go 19:31:16 <Rockyg> catherineD: Well, for users, maybe, but I think the project should test against multiple tags. At least the latest two. 19:31:17 <sslypushenko__> davidlenwell: You can find more info here: https://review.openstack.org/#/c/165921/ 19:31:41 <davidlenwell> sslypushenko__: I will review this and we can discuss later 19:32:34 <sslypushenko__> Sure... Join to discussion in review) 19:32:42 <davidlenwell> will do 19:32:58 <hogepodge> mtreinish: sslypushenko__ I left a comment on the patch, I think tagging classes would fix everything and keep compatibility. 19:33:09 <hogepodge> also, if refactoring is happening, tag for tempest should match test names. 19:33:36 <hogepodge> 2015.03 is a month old now, and 2015.04 was approved 19:33:47 <hogepodge> Board meeting last week. 19:33:57 <hogepodge> 2015.05 to be voted on in vanvouver. 19:34:02 <catherineD> Rockyg: we will need to discuss about that .. rightnow at installation time refstack-client install a certain tag (tag-3 for now)... so test results will be stable ... 19:34:22 <hogepodge> tag-3 will fail tests because of recent patches. 19:34:23 <sslypushenko__> hogepodge: mtreinish: I will propose next patch with uuid pairs soon 19:34:33 <Rockyg> OK. Also, hogepodge: +1 on tagging classes 19:35:21 <catherineD> so we agree to use tag-4, right? anyone disagree .. 19:36:14 <catherineD> #agreed upgrade refstack-client to tag-4 when it becomes available .. 19:36:15 <davidlenwell> So you are suggesting we use tag 4 before its offical release? 19:36:17 <sslypushenko__> catherineD agreed 19:37:02 <catherineD> davidlenwell: only when it is official release .. 19:37:12 <davidlenwell> okay .. then I agree 19:37:56 <catherineD> hogepodge: how about you? 19:39:03 <hogepodge> I would use whatever is tempest head as of now, but it's your call. 19:40:11 <davidlenwell> until we have complete uuid I'd avoid using trunk 19:40:16 <Rockyg> If tag-3 is already a problem, I would thing just top of master would be the right place at the moment. 19:40:20 <davidlenwell> once we do then I want trunk 19:40:43 <catherineD> hogepodge: since there are constant changes in tempest head ... the idea is to stabalize the test system itself ...so we think it would be better that all interops test are tested with same release .. 19:40:47 <hogepodge> tag-3 is hopelessly broken for our use case 19:40:56 <hogepodge> trunk isn't. 19:41:45 <hogepodge> git hash, uuid, test name, all uniquely identify a test. But again, your call. For defcore we're allowing trunk right now for test results 19:41:46 <catherineD> that is why we need to upgrade ... we know tag-3 needed upgrade ... the question here is to what? 19:42:08 <davidlenwell> hogepodge: is suggesting trunk 19:42:31 <davidlenwell> as long as we aren't still seeing a lot of path changes I don't have an issue with using trunk 19:42:32 <hogepodge> I would use 57ae3795 19:42:36 <catherineD> that mean the test system itself is not fixed ... 19:42:53 <catherineD> so I guess we do not agree to wait until tag-4? 19:43:03 <hogepodge> It's not my call. 19:43:06 <davidlenwell> wel using a specific git hash is the same as using a tag as far as I am concerned 19:43:28 <davidlenwell> if using 57ae3795 is more useful for defcore then lets use it until tag4 is done or uuid is finished 19:43:30 <catherineD> hogepodge: I see that you are a user of refstack ... since you run test with vendors. 19:43:31 <Rockyg> tag-4 will happen before Vancouver, but not for another few weeks. can we afford to wait? 19:43:57 <davidlenwell> we want all incoming data to be useful for defcore 19:44:16 <catherineD> davidlenwell: agree ... 19:44:18 <davidlenwell> if we are still onboarding testers before the summit then we should probably use what defcore is suggesting 19:44:42 <hogepodge> catherineD: I've been giving this (or a variant of it) to vendors for testing. https://review.openstack.org/#/c/167620/1/2015.03/2015.03.required.txt,cm 19:45:37 <catherineD> and what tempest realease will the vendor tested that with? head? 19:45:59 <hogepodge> Yes, because right now head works. tag-3 does not. 19:46:29 <hogepodge> (with flagged caveats) 19:46:49 <catherineD> how do we guarantee that the test path will not change? 19:46:55 <davidlenwell> I vote we go with hogepodge recomendations on this issue 19:47:26 <catherineD> from the tests in https://review.openstack.org/#/c/167620/1/2015.03/2015.03.required.txt,cm 19:47:33 <hogepodge> We don't. If the test path changes I'll give them a new file, traced with the UUID. I can also pin to a git hash from when before it changed. 19:48:38 <Rockyg> hogepodge: thanks for keeping up on this. Keeping the vendors testing is important... 19:48:46 <davidlenwell> I think our first priority right now is to be useful to defcore .. if the tag we insist on using isn't useful to them we kinda have a problem.. we can worry about making the data more useful over long term once we finish uuid .. for now lets focus on making sure defcore has what it needs 19:49:13 <catherineD> so I guess what we agree on now is to use tempest head? 19:49:24 <davidlenwell> yes .. I think so 19:49:27 <catherineD> ok 19:49:56 <davidlenwell> we need to follow through on uuid so that we can insure that our data is meaningful over time 19:49:57 <catherineD> I do I retrack the last agreed statement that I typed in because it is no longer true .. 19:50:12 <davidlenwell> just make anew one that says that 19:50:43 <catherineD> davidlenwell: I agreed there is no point for refstack to deviate from what the vendor is using ... 19:50:59 <catherineD> it would be a very easy change to refstack ... 19:51:47 <catherineD> #agreed refstack-client will be updated to use tempest head instead of tag release 19:51:49 <davidlenwell> agreed 19:51:58 <davidlenwell> this has been the goal all along 19:52:03 <catherineD> ok moving on .. 19:52:10 <davidlenwell> you might remember me ranting about this almost 2 years ago 19:52:52 <davidlenwell> Do we have anything else to discuss in this meeting? we are running short on time? 19:53:09 <catherineD> #topic pending review 19:54:26 <catherineD> sslypushenko__: could you review https://review.openstack.org/#/c/168512/ 19:54:48 <sslypushenko__> Sure... 19:55:28 <catherineD> the only concern I have on this patch is how do we make sure that the capability file is up-to-date ... 19:55:47 <catherineD> since the file is now in defcore reppsitory ... 19:55:50 <davidlenwell> catherineD: and I had a discussion about this offline 19:56:08 <davidlenwell> I was thinking that we should have an api method that keeps a cached copy from the repo 19:56:35 <davidlenwell> that way our front end is only ever making calls to our api 19:56:42 <catherineD> davidlenwell: +1 19:57:05 <davidlenwell> sslypushenko__: what do you think? 19:57:33 <sslypushenko__> agreed 19:57:40 <davidlenwell> okay 19:57:41 <catherineD> great 19:57:46 <davidlenwell> done .. and just in time 19:58:04 <davidlenwell> lets clear out .. the next group likes to start on time 19:58:05 <catherineD> I am testing sslypushenko__: 's patch 19:58:26 <catherineD> I that is all the time we have today .... 19:58:45 <catherineD> any other discussion in 2 mins ? 19:59:11 <catherineD> thank you all ... 19:59:19 <catherineD> #endmeeting