14:09:33 <boris-42> #startmeeting Rally
14:09:34 <openstack> Meeting started Mon Jun  8 14:09:33 2015 UTC and is due to finish in 60 minutes.  The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:09:35 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:09:38 <openstack> The meeting name has been set to 'rally'
14:09:38 <boris-42> yfried: ping
14:09:42 <boris-42> redixin: ping
14:09:45 <boris-42> amaretskiy: ping
14:09:51 <amaretskiy> hi
14:09:53 <msdubov_> hi
14:09:56 <boris-42> msdubov_: hi
14:09:57 <yfried> hi
14:10:50 <skraynev> hi
14:11:20 <boris-42> skraynev: hi there
14:11:24 <e0ne> hi
14:11:27 <boris-42> stpierre: ping
14:11:31 <stpierre> pong
14:12:15 <boris-42> so let's start=)
14:12:34 <boris-42> #topic Plugin base finally=)
14:12:56 <boris-42> https://review.openstack.org/#/c/160142/10
14:13:03 <boris-42> ^ more or less I finished it
14:13:20 <boris-42> one more bonus that it adds (except removal of copy pasted code & unfification of plugins)
14:13:31 <boris-42> is that we will get namespaces for plugins
14:13:36 <yfried> boris-42: great
14:13:49 <boris-42> which is crucial if we are going to benchmark other ecosystems
14:13:49 <yfried> boris-42: although pboldin seems to disagree :)
14:13:53 <boris-42> e.g. doker
14:14:01 <boris-42> yfried: it's okay=)
14:14:44 <boris-42> so this is actually one of blocker for cutting Rally 0.1.0 release
14:14:52 <boris-42> so please review it!
14:15:03 <boris-42> okay let's move to next topic
14:15:13 <boris-42> #topic Rally core structure changes
14:15:30 <boris-42> We don't scale enough and if we add more projects this will be even worse
14:15:46 <boris-42> so I introduced 2 types of cores
14:16:01 <boris-42> cores that are reviewing everything
14:16:16 <boris-42> and cores that are reviewing plugins that are specific for their project
14:16:26 <boris-42> (usually cores of those projects)
14:16:36 <yfried> boris-42: is this enforced by gerrit, or just good will?
14:16:54 <boris-42> yfried: sooo. I know how to enforce it by gerrit
14:17:13 <boris-42> yfried: but community was against and they said that I am very bad person and rally project sux=)
14:17:22 <boris-42> yfried: and I should trust people
14:17:30 <yfried> boris-42: let's not get into this.
14:17:42 <boris-42> yfried: so we will just use Trust model
14:17:51 <boris-42> yfried: I will keep up-to-date this page https://rally.readthedocs.org/en/latest/project_info.html
14:18:00 <yfried> boris-42: do you have the auth to set gerrit as you suggested
14:18:02 <yfried> ?
14:18:19 <boris-42> yfried: I am not allowed to touch gerrit
14:18:37 <boris-42> yfried: as I said I proposed this to mailing list...
14:18:46 <yfried> boris-42: why not go with the +A module accepted in Neutron? seems like it's a halfway solution
14:18:55 <yfried> boris-42: I've read the thread.
14:19:03 <boris-42> yfried: I don't like trust model
14:19:11 <boris-42> yfried: if we use trust model there is no difference
14:19:11 <yfried> boris-42: I agree
14:19:19 <boris-42> yfried: between neutron stuff and our
14:19:29 <boris-42> yfried: we just trust cores from other project to maintain their plugins
14:19:42 <boris-42> yfried: and they just said that they won't +2 patches not related to their topic
14:20:51 <boris-42> yfried: I think overall it's good, just sad that it's not automated
14:20:56 <yfried> boris-42: so , IIUC, the Neutorn module says "leutantnats" can +2 but not +A
14:21:42 <boris-42> yfried: the same here
14:21:56 <yfried> boris-42: ok. so we go with trust.
14:21:57 <boris-42> yfried: a bit actually better if full core put +2
14:22:06 <boris-42> yfried: they can +2+A
14:22:35 <yfried> boris-42: let's just have it on record that patches can't touch plugins and "core" at the same time
14:22:40 <yfried> boris-42: what do you think?
14:22:57 <boris-42> yfried: this will be enforced
14:23:04 <boris-42> yfried: by nerd reviewer some day
14:23:20 <yfried> boris-42: for now, can it be in doc as policy?
14:23:45 <boris-42> yfried: so where you would like to put this?
14:24:05 <yfried> boris-42: don't know. isn't msdubov_ our doc guru?
14:24:24 <boris-42> yfried: heh
14:24:30 <yfried> boris-42: the same place we state that you can't reduce coverage
14:24:33 <boris-42> yfried: so probably just in project_info
14:24:45 <boris-42> yfried: reducing coverage is automated IMHO
14:24:55 <boris-42> yfried: like it won't just pass CI
14:25:00 <yfried> boris-42: doesn't meen it shouldn't be written in doc
14:25:28 <yfried> boris-42: new contrib should be able to read all the guidlines
14:25:42 <boris-42> yfried: hm we have testing guidlines
14:25:57 <boris-42> yfried: https://github.com/openstack/rally/tree/master/tests
14:26:46 <boris-42> yfried: but not sure that passing CI should be documented
14:26:53 <boris-42> yfried: it's obvious, no?
14:27:55 <yfried> boris-42: the TYPE of CI should be doc. so ppl can easily understand what the job is testing and what they are required to do. but we are digressing. let's go back to mtg
14:28:32 <boris-42> yfried: in any case description of coverage job should go here
14:28:37 <boris-42> yfried: https://github.com/openstack/rally/tree/master/tests
14:28:51 <boris-42> yfried: and non touching framework & plugins should go somehwere else
14:30:02 <boris-42> okay let's move to next topic
14:30:14 <boris-42> #topic Rally as a Lib & Rally as a Service (finally)
14:30:19 <boris-42> msdubov_: ping
14:30:30 <boris-42> msdubov_: could you share efforts?
14:30:55 <msdubov_> boris-42: sure
14:30:57 <msdubov_> #link https://review.openstack.org/#/c/182245/
14:31:08 <msdubov_> For now, we have a spec describing future changes ^
14:31:42 <msdubov_> The spec includes several code snippets describing the way we're going to implement RaaS to make it simple to use Rally both in local and in RPC mode
14:32:05 <yfried> msdubov_: nice
14:32:11 <msdubov_> There are also several proposals on changing the API itself, to move all DB/object calls there, to make CLI thin etc.
14:32:24 <msdubov_> yfried: Please review it!
14:32:28 <yfried> msdubov_: will do
14:32:39 <msdubov_> eom
14:32:43 <boris-42> msdubov_: nice work
14:32:49 <boris-42> msdubov_: i will review it as well soon
14:32:58 <boris-42> msdubov_: there is a plenty amount of work to do
14:33:13 <boris-42> msdubov_: I believe we will need to cut 0.2.0 to get all these changes in
14:33:22 <msdubov_> boris-42: Maybe
14:33:36 <boris-42> msdubov_: seems like not maybe=)
14:34:32 <boris-42> okay let's move to next topic
14:34:37 <boris-42> #topic Infra updates
14:34:50 <boris-42> redixin: please could you share all thing that you are working on now?
14:36:36 <redixin> im trying to speed up testing change requests in rally
14:36:37 <boris-42> redixin: around?
14:37:58 <redixin> we split one big rally job to 3 small jobs. we need to merge one patch https://review.openstack.org/#/c/185031/
14:38:28 <redixin> because now we only testing only ~33% of scenarios =)
14:38:44 <boris-42> redixin: 33% ?
14:38:52 <boris-42> redixin: what?)
14:39:21 <redixin> oh no =)
14:39:52 <redixin> we testing all scenarios, but in one job
14:40:21 <redixin> anyway after merging that patch we still have one slow job
14:40:28 <redixin> rally-verify
14:40:46 <redixin> and rally-cli
14:41:28 <redixin> two slow jobs
14:41:34 <boris-42> redixin: so rally-veriy will be refactored
14:41:39 <e0ne> boris-42, redixin: we need to add more scenarios to gates for openstack components
14:42:00 <boris-42> redixin: andreykurilin is working on it
14:42:25 <boris-42> redixin: rally-dsvm-cli I don't know how to refactor it, could we get at least timings of tests?
14:42:29 <redixin> boris-42: rally-verify will be splitted too?
14:42:30 <boris-42> redixin: so we will know the slowest one?
14:42:33 <boris-42> redixin: nope
14:42:45 <boris-42> redixin: just we will run not so big tempest sets
14:43:09 <boris-42> e0ne: so now we have more jobs
14:43:17 <redixin> so we have only unsolved slow tox -ecli
14:43:38 <boris-42> redixin: yep we need to find the way to display duration
14:43:57 <redixin> e0ne currently we have jobs for heat, nova, neutron and one jobs for the rest
14:44:03 <e0ne> boris-42: we add new scenarios to rally. e.g. in cinder repo we don't have all of cinder scenarios:(. my fail
14:44:31 <boris-42> e0ne: so it's not big deal
14:44:31 <redixin> e0ne feel free to add new sceanarios to appropriate file in rally-jobs/
14:44:40 <boris-42> e0ne: yep
14:45:12 <e0ne> boris-42, redixin: imo, need to add a note about it to the contibutor's guide
14:45:34 <boris-42> e0ne: ?
14:45:48 <redixin> e0ne +1
14:45:49 <boris-42> e0ne: you mean to add note in rally project to add in jobs of cinder?
14:46:02 <e0ne> boris-42: yep
14:46:14 <boris-42> e0ne: it's not enough to have readme in cidner?
14:46:17 <boris-42> e0ne: https://github.com/openstack/cinder/tree/master/rally-jobs ?
14:46:18 <e0ne> personally, i forget to do it almost evety time:(
14:47:00 <redixin> boris-42: how do people know where to add newly created scenarios?
14:47:02 <e0ne> boris-42: the problem is that people can contribute to rally and don't read readme in cinder or nova or heat or etc:(
14:47:27 <boris-42> redixin: e0ne and where we should add such readme?
14:47:45 <boris-42> here https://github.com/openstack/rally/tree/master/rally/plugins/openstack ?
14:47:48 <boris-42> e0ne: ^
14:47:51 <redixin> boris-42: is there a "rally contributor guide"?
14:48:18 <boris-42> redixin: there is
14:48:27 <boris-42> redixin: https://rally.readthedocs.org/en/latest/contribute.html
14:48:41 <boris-42> redixin: not sure that anybody is reading it twice
14:48:51 <boris-42> I believe that we should put here small noet
14:48:53 <boris-42> note*
14:48:53 <boris-42> https://rally.readthedocs.org/en/latest/contribute.html
14:49:00 <boris-42> https://github.com/openstack/rally/tree/master/rally/plugins/openstack  **
14:49:03 <boris-42> here ^
14:49:09 <e0ne> boris-42: +1
14:50:16 <boris-42> e0ne: be free to add it lol=)
14:50:19 <boris-42> e0ne: your idea=)
14:50:26 <e0ne> hm...
14:50:40 <boris-42> e0ne: =))
14:50:56 <boris-42> e0ne: okay someday somebody will add it
14:51:18 <e0ne> boris-42: sounds like not a good plan
14:53:43 <boris-42> e0ne: no worries=)
14:53:55 <boris-42> e0ne: I am going to cleanup all Readme files
14:54:05 <boris-42> during that cleanup I will add some note as well in rally/plugins/*
14:56:37 <boris-42> okay
14:56:41 <boris-42> #topic free disucssion
14:56:47 <boris-42> Does anybody has to say anything?
14:58:09 <boris-42> Thank you for coming and seems like we are out of time. So see you next week
14:58:13 <boris-42> #endmeeting