14:00:49 #startmeeting Rally 14:00:49 Meeting started Mon Jul 6 14:00:49 2015 UTC and is due to finish in 60 minutes. The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:50 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:53 The meeting name has been set to 'rally' 14:01:12 yfried: stpierre ping 14:01:17 pong 14:01:33 boris-42: hi 14:02:04 hi all 14:02:30 skraynev: hi there 14:02:40 e0ne: hi 14:02:44 hi all 14:02:50 o/ 14:02:56 \o 14:02:57 boris-42: try to split mind on two sync-ups :) 14:03:06 amaretskiy: ping 14:03:08 hi 14:03:15 \o 14:03:15 yingjun: hi hi 14:03:23 hi 14:03:38 temujin: o//////// 14:03:47 pffff 14:04:10 Hi 14:04:31 boris-42: hi 14:04:53 okay let's get start) 14:05:14 #topic Rally Meeting Agenda (finally we did it) 14:05:32 rvasilets: Thank you for you hard work on this not so interesting task 14:05:49 boris-42: please post a link 14:05:52 rvasilets: could you explain how it is organized ? and what you did 14:06:13 boris-42, sure 14:06:31 * yfried will be back in 10min 14:06:46 We could find a link in priject info in wiki Rally page 14:06:48 https://wiki.openstack.org/wiki/Rally#Project_Info 14:06:59 there is Project meetings 14:07:04 https://wiki.openstack.org/wiki/Meetings/Rally 14:07:36 #link https://wiki.openstack.org/wiki/Meetings/Rally 14:08:05 Before every weekly meeting we should post our suggestion to agenda. After a meeting I will briefly summarize our meeting in Previous Meetings 14:08:26 eom 14:08:47 rvasilets: as well we should add that meeting manager will be changed each 2 weeks 14:09:11 rvasilets: so nobody will get to tired from doing this work 14:09:17 rvasilets: why you don't use meeting notes for summary? 14:09:29 e0ne: because it's not human written 14:09:35 :) 14:09:48 e0ne: the same why we don't use release notes that are auto generated 14:10:12 good point:) 14:10:24 e0ne: the same why we don't use tempest for testing rally 14:11:03 andreykurilin: LOL 14:11:10 okay so the idea is next 14:11:25 if you want to discuss something you should add it to agenda 14:11:51 in such way we will have more organized and useful meeting 14:12:08 boris-42: +! 14:12:10 boris-42: +1 14:12:27 skraynev: yingjun stpierre redixin ^ 14:12:53 boris-42: ok. currently is nothing from me 14:13:06 looks good 14:13:21 I thought it is already like that ^ 14:14:03 redixin: hehe 14:14:48 btw, IMO we should try to use all this meetbot stuff, and not bother people with making meeting summary by hand 14:15:45 redixin: something like: 14:15:47 redixin: this is good point to have better logs of meeting (but I would still prefer to have human written summaries) 14:15:50 #idea we should try to use all this meetbot stuff, and not bother people with making meeting summary by hand 14:15:54 ? 14:16:05 andreykurilin, yep 14:16:15 #help 14:16:25 https://wiki.debian.org/MeetBot 14:16:31 #link https://wiki.debian.org/MeetBot 14:17:41 yhis bot is very limited 14:17:51 bounded 14:18:15 I don't now what tag to use when I whant to summerazi the topic 14:18:18 for example 14:18:28 * now=>know 14:18:47 *summarize 14:18:50 #agreed? 14:19:07 ) 14:19:17 #agreed We should use https://wiki.debian.org/MeetBot tags 14:19:27 ok its good idea 14:19:39 don;t forget to use them then! 14:19:41 #agreed We should write summaries by hand 14:19:57 #agreed we should rotate each 2 weeks meeting manager 14:20:14 #action We should create list of people who would like to help with task 14:20:28 okay let's move to the first topic 14:20:30 40 minutes reminder :) 14:20:40 #topic Ironic benchmarks 14:20:50 rvasilets: so? 14:21:59 okey. I tried to add first Ironic benchmark https://review.openstack.org/#/c/186064/ and I thought that its done, but... 14:22:30 I found some problem that I cant solve 14:22:38 http://logs.openstack.org/64/186064/14/experimental/gate-rally-dsvm-ironic-rally/9b0a8d7/console.html#_2015-06-30_14_39_09_363 14:23:00 I got recursion and I don't know why! 14:23:16 Need help 14:23:32 This patch is desired for Ironic team to enable Rally job in Ironic tests 14:23:36 rvasilets: wrong link 14:23:37 eom 14:23:52 http://logs.openstack.org/64/186064/14/experimental/gate-rally-dsvm-ironic-rally/9b0a8d7/console.html.gz#_2015-06-30_14_39_09_363 rvasilets this one shoudl work 14:23:57 btw you should use 14:23:59 #link http://logs.openstack.org/64/186064/14/experimental/gate-rally-dsvm-ironic-rally/9b0a8d7/console.html.gz#_2015-06-30_14_39_09_363 14:24:43 rvasilets: btw you forgot to send link to the patch actually* 14:24:46 okey. Sorry for that. The link is #link https://review.openstack.org/#/c/186064/ for the patch that need help 14:25:44 Is there someone who can help me with that? 14:26:06 #help Is there someone who can help me with that? 14:26:19 rvasilets: so did you take a look at the code of python ironic client? 14:26:28 rvasilets: did you try to run by hands this commands and so on? 14:26:35 yes 14:26:42 rvasilets: and they work? 14:27:08 cli of Ironic is working 14:27:30 rvasilets: not CLI 14:27:40 rvasilets: did you use it as python client (lib) 14:27:53 rvasilets: in exactly the same way as rally use it 14:28:25 no 14:28:45 rvasilets: so try it firstly (to repeat the same commands) 14:29:20 rvasilets: if they work then there is some difference between what you did in rally and by hand 14:29:38 rvasilets: if they doesn't work there is difference between CLI and what you are doing 14:29:45 rvasilets: in any case I will try to take a look at this 14:30:08 But there is one more step to find that difference! 14:30:24 #action rvasilets will try to understand why Ironic benchmarks doesn't work and why CLI of ironic works 14:30:44 #action boris-42 will review Ironic patch https://review.openstack.org/#/c/186064/ 14:30:50 okay let's move to next topic 14:31:11 #topic Use DDT framework in tests 14:31:28 Okay guys let's use DDT framework for testing and rewrite all the tests on it 14:31:41 This framework is really nice 14:31:47 #idea tag?) 14:31:53 we can not use ddt in all tests - only where that is usefult 14:31:57 rvasilets: nope because it's topic 14:32:13 amaretskiy: yep but there is a lot of places where it is useful =) 14:32:20 vponomaryov: ^ 14:32:30 boris-42: ok :) 14:32:30 amaretskiy: part of group of tests can be rewritten on it 14:32:33 boris-42: what is DDT? what's its adv? 14:32:40 this http://ddt.readthedocs.org/en/latest/ ? 14:32:41 yfried: let me show the sample 14:33:20 yfried: https://github.com/openstack/rally/blob/master/tests/unit/plugins/openstack/scenarios/manila/test_utils.py#L117-L133 14:33:30 #link sample of DDT way to test https://github.com/openstack/rally/blob/master/tests/unit/plugins/openstack/scenarios/manila/test_utils.py#L117-L133 14:34:08 so?) 14:34:43 this module btw can be replaced with DDT https://github.com/openstack/rally/blob/97c520c828fe7d44fb6b0ce3011bc23eefba1bb1/tests/unit/plugins/openstack/scenarios/neutron/test_loadbalancer_v1.py#L100-L127 14:35:13 boris-42: 100% 14:35:14 seems like in some cases it'd be very useful, and in others can provide some marginal benefit by easily multiplying test cases (and thus hopefully increasing edge-case coverage) 14:35:35 but i don't think we need to be quite so dogmatic as rewriting *all* the tests with ddt :) 14:35:57 * yfried agrees with stpierre 14:36:13 but it'd certainly simplify any of the cases where we have a _test_foo() function, and that's called from test_foo() and test_foo_failure() and test_foo_with_bar() and .... 14:36:15 stpierre: yfried rewritting *all* the tests that should be rewriten 14:36:22 we have enough on our plate without rewriting existing tests 14:36:36 boris-42: rally coverage seems excellent already 14:36:48 yfried: house keeping is very important task 14:37:03 yfried: I mean it looks like not important task, but it's actually curcial 14:37:06 boris-42: I don't disagree, it's just priorities... 14:37:16 we should definitely keep ddt in mind for code reviews, and it'd be nice to rewrite obvious low-hanging fruit so that we have examples to point at 14:37:26 stpierre: yep 14:37:36 stpierre: these tasks are quite good for newbies 14:37:46 stpierre: what we need is just set of bugs 14:37:47 and some of us seem to have a particular masochistic bent for working on unit tests 14:37:59 stpierre: LOL 14:38:16 boris-42: stpierre: I guess if someone volunteered to flag all refactor candidates and tag them as low-hanging fruits we can pick it up later 14:38:52 #info yfried: so what we need for now is to block intorducing new cases where DDT should be used 14:39:00 #agree using ddt for newly added tests, low priority to rewite 14:39:26 everybody agree ^ ? 14:39:32 #agree 14:39:38 boris-42: I'll have to look into DDT to know this for blocking new patches. could you please provide a good link? 14:40:01 yfried: so this thing is quite simple 14:40:16 it's just decorator that creates N times the same method with different arguments 14:40:30 I belive the best sample is to take a look at manila plugins tests 14:41:41 boris-42: ack 14:42:16 why don't we try to get a couple of existing tests converted as well, then we'll have a larger body of examples 14:42:37 stpierre: good #idea 14:42:44 we don't need to do them all right now, but a couple more would be nice 14:43:10 stpierre: yfried btw kiran can refactor 14:43:17 stpierre: yfried own tests to use DDT 14:43:42 and we will have a sample of patch that refactors part of tests 14:43:44 boris-42: if kiran is willing to be the guinee pig :) 14:43:45 that we can show to other 14:44:11 yfried: I think he likes to learn new stuff 14:45:29 okay let's move to next topic 14:45:58 #topic Optimize Rally imports 14:46:07 Okay guys I did everything that was possible 14:46:16 (without changing other libs) 14:46:32 #link https://review.openstack.org/#/c/196493/ and here is the result of my work 14:47:08 yfried: ^ 14:47:50 boris-42: +2 14:48:18 boris-42: it's a good patch 14:48:31 The only thing that I am worried is that we are able to introduce regression in future 14:48:55 but there are part of really good changes e.g. lazy db that saves most of the time=) 14:49:29 okay let's move to next topic 14:49:37 #agree That patch should be merged =) 14:49:43 #agreed That patch should be merged =) 14:49:52 #topic Some review work update 14:49:54 kun_huang: ping 14:50:04 boris-42: pong 14:50:14 there some patches need to track 14:50:21 https://etherpad.openstack.org/p/rally_review_work 14:51:05 the one who submit sahara related patches seems leaving 14:51:06 kun_huang: why do we need this? https://review.openstack.org/#/dashboard/?foreach=%28project%3Aopenstack%2Frally%29+status%3Aopen&title=Rally+Dashboard&Critical+for+next+release=%28starredby%3A%22mdubov%40mirantis.com%22+AND+starredby%3A%22boris%40pavlovic.me%22%29+AND+status%3Aopen&Waiting+for+final+approve=label%3AVerified%3E%3D1%252cjenkins+label%3AVerified%3E%3D1%252crally%252Dci+NOT+owner%3Aself+label%3ACode%252DReview% 14:51:07 3E%3D2+NOT+label%3ACode%252DReview%252D1+NOT+label%3AWorkflow%3E%3D1+NOT+label%3AWorkflow%3C%3D%252D1+NOT+owner%3Aself+label%3ACode%252DReview%3E%3D2+NOT+label%3ACode%252DReview%252D1+NOT+label%3AWorkflow%3E%3D1+NOT+label%3AWorkflow%3C%3D%252D1+NOT+%28starredby%3A%22amaretskiy%40mirantis.com%22+OR+starredby%3A%22akurilin%40mirantis.com%22+OR+starredby%3A%22boris%40pavlovic.me%22+OR+starredby%3A%22yingjun.li%40kylin%252Dcloud. 14:51:07 com%22+OR+starredby%3A%22mdubov%40mirantis.com%22+OR+starredby%3A%22sskripnick%40mirantis.com%22+OR+starredby%3A%22yfried%40redhat.com%22%29&Bug+fixes=topic%3A%22%5Ebug%2F.%2A%22+AND+status%3Aopen&Proposed+specs=%28message%3A%22%5Bspec%5D%22+file%3A%22%5E.%2Aspecs.%2A%22%29&Important+patches=%28starredby%3A%22boris%40pavlovic.me%22+OR+starredby%3A%22mdubov%40mirantis.com%22%29+NOT+%28starredby%3A%22mdubov%40mirantis.com% 14:51:07 22+AND+starredby%3A%22boris%40pavlovic.me%22%29+AND+status%3Aopen&Ready+for+review=label%3AVerified%3E%3D1%252cjenkins+label%3AVerified%3E%3D1%252crally%252Dci+NOT+label%3ACode%252DReview%252D1+NOT+label%3AWorkflow%3E%3D1+NOT+label%3AWorkflow%3C%3D%252D1+NOT+label%3ACode%252DReview%3E%3D2+NOT+%28starredby%3A%22amaretskiy%40mirantis.com%22+OR+starredby%3A%22akurilin%40mirantis.com%22+OR+starredby%3A%22boris%40pavlovic.me% 14:51:07 22+OR+starredby%3A%22yingjun.li%40kylin%252Dcloud.com%22+OR+starredby%3A%22mdubov%40mirantis.com%22+OR+starredby%3A%22sskripnick%40mirantis.com%22+OR+starredby%3A%22yfried%40redhat.com%22%29&Has+%252D1+but+passed+tests=label%3AVerified%3E%3D1%252cjenkins+label%3AVerified%3E%3D1%252crally%252Dci+label%3ACode%252DReview%252D1+NOT+label%3AWorkflow%3E%3D1+NOT+label%3AWorkflow%3C%3D%252D1+NOT+label%3ACode%252DReview%3E%3D2 14:51:08 sorry 14:51:20 yfried: uh oh=) 14:51:36 yfried: I know this, it not 100% enough 14:51:37 goo.gl/04aT6p 14:51:43 boris-42: kun_huang: tried to post the link to boris-42's nice board 14:51:56 yfried: here is google short url goo.gl/04aT6p 14:52:15 kun_huang: so maybe we can extend that dashboard? 14:52:31 boris-42: yep, I will try it 14:52:36 kun_huang: .e.g. if I mark patch with star it will become important one 14:52:52 kun_huang: if it has +2 it is close to merge (e.g. in separated column) 14:53:03 kun_huang: we can add you to the people who can manage important patches 14:53:19 kun_huang: if you would like to maintain this kind of work 14:53:27 boris-42: :-) 14:54:10 kun_huang: so what do you think? 14:54:38 boris-42: “people who can manage important patches” means core? 14:54:50 kun_huang: nope it's all by hand ... 14:55:02 kun_huang: btw I belive we can put acl group 14:55:10 kun_huang: or create new one specially for this 14:55:17 kun_huang: I will try to play with it 14:55:38 boris-42: me too, I would love to have a try on it 14:56:31 let's go back to my topic it self first 14:56:33 #action Do the changes with rally dashboard to fit kun_huang requirments) 14:56:42 some came to rally and contribute their patches 14:56:44 kun_huang: we have only 4 minute s=) 14:56:51 kun_huang: so be fast=) 14:57:08 some of patches seems be ignored and they are leaving 14:57:57 (my network broken just now) 14:58:28 kun_huang: so we are trying not to ignore stuff 14:58:31 kun_huang: at least me... 14:59:05 kun_huang: so in any case we should add new rule 14:59:20 #idea e.g. review the oldest patches firstly * 14:59:21 boris-42: kun_huang: I believe the dashboard should be able to show verifed patches that haven't been reivewd for a long while 14:59:32 yfried: yep exactly 14:59:43 good idea 14:59:48 #agree - review the oldest patches firstly 14:59:54 kun_huang: we will try to add some kind of sort 15:00:00 fifo 15:00:02 like latest review from core 15:00:05 #endmeeting