14:02:45 <rvasilets_> #startmeeting Rally
14:02:46 <openstack> Meeting started Mon Jul 11 14:02:45 2016 UTC and is due to finish in 60 minutes.  The chair is rvasilets_. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:48 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:02:50 <openstack> The meeting name has been set to 'rally'
14:03:31 <rvasilets_> amaretskiy, ikhudoshyn andreykurilin__  boris-42
14:03:34 <rvasilets_> hi ^
14:03:35 <amaretskiy> hi
14:03:53 <ikhudoshyn> o/
14:04:05 <andreykurilin> o/
14:04:28 <boris-42> hi hi
14:04:39 <rvasilets_> Okey looks like almost all are here
14:04:59 <rvasilets_> #topic [amaretskiy] Rally-as-a-Service demonstration
14:05:02 <rvasilets_> [amaretskiy] We have Rally-as-a-Service demonstration - let's discuss it and decide 1) if this demo implementation is good? and 2) do we need a spec based on this demo?
14:05:16 <rvasilets_> amaretskiy, words to you
14:05:20 <amaretskiy> #link https://review.openstack.org/#/c/336636/
14:05:46 <amaretskiy> The commit message worth all my words :) please take a look to it
14:06:11 <amaretskiy> I wish you review this patch!!!
14:06:57 <amaretskiy> and give feedback if this implementation is good and we can proceed to have a merge-candidate patch
14:07:09 <amaretskiy> and, if required, a spec for RaaS
14:07:32 <amaretskiy> eom
14:07:53 <boris-42> amaretskiy: I haven't chance yet
14:08:02 <boris-42> amaretskiy: however I like the amount of the code
14:08:03 <boris-42> =)
14:08:10 <amaretskiy> great!!!!!!!
14:08:16 <rvasilets_> Oh, this is the huge patch and I beliave its important to review it as quick as possible)
14:09:02 <amaretskiy> the major part of this patch is supporting code that is not a body of proposal, so real amount of code is even lesser
14:09:29 <rvasilets_> Do we have blockers for that patch?
14:10:08 <amaretskiy> we do not have blockers
14:10:48 <amaretskiy> we could decide if we need a Spec
14:10:58 <amaretskiy> boris-42 ^
14:11:20 <boris-42> amaretskiy: so I don't think that we need a spec
14:11:27 <boris-42> amaretskiy: let's keep it on review for now
14:11:28 <amaretskiy> great
14:12:02 <rvasilets_> Okey any questions for that topic? ikhudoshyn andreykurilin
14:12:06 <andreykurilin> no
14:12:07 <ikhudoshyn> we don't want to publish an API for RaaS right
14:12:08 <ikhudoshyn> ?
14:12:21 <ikhudoshyn> only our client <-> our service?
14:12:34 <andreykurilin> hm..
14:12:51 <rvasilets_> I thought we should...
14:13:03 <ikhudoshyn> otherwise we'd better have a spec/doc for the API to make sure it is designed properly
14:13:17 <amaretskiy> ikhudoshin, i do not understand the question
14:13:27 <ikhudoshyn> changing existing api after it gets published would be such a pain
14:13:31 <boris-42> ikhudoshyn: actually we are going to publish
14:13:58 <ikhudoshyn> boris-42: so we'd better have it (the API) done right from the 1st attempt
14:14:01 <boris-42> ikhudoshyn: we can call it actually experimental
14:14:08 <boris-42> ikhudoshyn: and change in any way
14:14:13 <ikhudoshyn> sounds reasonable
14:14:19 <amaretskiy> ikhudoshyn: RaaS API is something stable and should be unchanged/versioned
14:14:24 <ikhudoshyn> ok i'm fine with that
14:14:57 <amaretskiy> we could change public API during the beta period
14:15:28 <ikhudoshyn> well it SHOULD be stable and  unchanged/versioned
14:15:34 <rvasilets_> Okey then I believe we are #agreed To review Raas patch
14:15:35 <amaretskiy> sure
14:15:47 <ikhudoshyn> yet supporting multiple versions is not an easy task
14:16:08 <ikhudoshyn> so i'm just about be sure that we're absolutely ok with the API
14:16:10 <boris-42> ikhudoshyn: amaretskiy so it is going to be expertimental
14:16:13 <boris-42> for a while
14:16:27 <boris-42> plus we are going to review it in patch
14:16:28 <ikhudoshyn> as boris-42 said we could keep it 'experimental' for a while
14:16:42 <ikhudoshyn> yup
14:16:49 <boris-42> it is better then reviewing it in spec
14:16:50 <boris-42> =)
14:16:57 <ikhudoshyn> (y)
14:17:05 <amaretskiy> ikhudoshyn: sure. Ut's enough to add an api_version parameter to RaaS
14:17:12 <amaretskiy> *It's
14:17:32 <ikhudoshyn> amaretskiy: )) you're straight as a ruler )
14:17:54 <amaretskiy> KISS-way :)
14:18:02 <ikhudoshyn> it's enough indeed) but it'd be better to avoid unnesecary versions at all
14:18:11 <amaretskiy> sure
14:18:23 * ikhudoshyn thub's up
14:18:25 <andreykurilin> btw, why we need public API? why not leave compatibility at rally-as-a-python-lib layer ?
14:18:42 <ikhudoshyn> andreykurilin: that was my initial Q
14:19:02 <andreykurilin> :)
14:19:04 <boris-42> andreykurilin: it would be good to allow people to work with rally via "curl"
14:19:13 <andreykurilin> oh...
14:19:19 <amaretskiy> agree 100% :)
14:19:28 <boris-42> so they won't need to install rally to use rally service
14:19:29 <ikhudoshyn> "API first"
14:19:30 <ikhudoshyn> ))
14:20:42 <andreykurilin> boris-42: imo, it is better to split rally to rally-api and rallyclient to decrease requirements of rallyclient and suggest to install rallyclient allways
14:21:41 <andreykurilin> *rally-api will provide both api and client styff; rallyclient will be just sync of code from rally-api project
14:22:04 <boris-42> andreykurilin: there is no rally client
14:22:06 <rvasilets_> Split rally?) Looks like openstack way)
14:22:09 <boris-42> andreykurilin: there is rally
14:22:30 <andreykurilin> I dislike curl approach at all
14:22:31 <andreykurilin> :)
14:22:36 <boris-42> andreykurilin: spliting rally and rally client will be nightmare
14:22:48 <amaretskiy> andreykurilin: "curl" meand urllib/requests/etc....
14:22:52 <amaretskiy> *means
14:22:52 <ikhudoshyn> boris-42: if there's no rally client then how does one gonna use RaaS?
14:22:53 <andreykurilin> i got it
14:23:13 <andreykurilin> amaretskiy: i got what curl means:)
14:23:22 <amaretskiy> even web browser
14:23:24 <boris-42> ikhudoshyn: so there is only one bash implmenetation of client
14:23:47 <boris-42> ikhudoshyn: no matter what you are using local/remote rally
14:23:57 <ikhudoshyn> boris-42: "pip install rally" just to connect to RaaS?
14:24:14 <andreykurilin> boris-42: it should not be a nightmare. just simple script which will copy files from rally.cmd and rally.api to separate repository
14:24:14 <boris-42> ikhudoshyn: and as amaretskiy said we won't be able to create web tier without API that is accessable via HTTP
14:24:25 <ikhudoshyn> i'm personally fine with that but people won't get it
14:24:25 <boris-42> ikhudoshyn: or use local
14:24:36 <boris-42> andreykurilin: think a bit more
14:24:55 <andreykurilin> I'm not saying that it is a good idea:) but I dislike support of public api:)
14:25:06 <boris-42> andreykurilin: you have to support it
14:25:11 <ikhudoshyn> andreykurilin: people love API
14:25:13 <boris-42> andreykurilin: otherwise no web UI
14:25:29 <boris-42> andreykurilin: integration will look like crap
14:25:42 <boris-42> andreykurilin: and how you are going to do smooth changes in api
14:25:49 <andreykurilin> so we need to start thinking about versioning, publishing our API and testing based on "curl"
14:26:00 <ikhudoshyn> andreykurilin: sure
14:26:02 <boris-42> because client/server code are on different servers
14:26:32 <boris-42> .kl
14:27:05 <andreykurilin> and we need to start doing it right now, since it is a hard topic
14:27:54 <boris-42> andreykurilin: amaretskiy seems like we need spec
14:28:05 <amaretskiy> if rally-client means another repository/project/pip-package then I would vote No
14:28:19 <boris-42> andreykurilin: I was already thinking about this splitting rally and rally client
14:28:19 <amaretskiy> boris-42: ok
14:28:24 <boris-42> amaretskiy: and it was shitty idea =)
14:28:30 <boris-42> andreykurilin: ^
14:28:33 <amaretskiy> but spec requires your reviews
14:28:33 <andreykurilin> as soon as we start provide public API, even if it experimental, we will unable to change it. We will have +100500 bug reports, since R-a-a-S is a hot topic
14:28:37 <amaretskiy> of demo
14:28:55 <andreykurilin> and everybody wants it
14:28:55 <boris-42> it didn't work for us because Rally should have only one CLI for both Rally local installation & Rally remove installation
14:28:57 <amaretskiy> please review demo and let's decide if proposal is good for spec
14:29:22 <boris-42> andreykurilin: otherwise it will be just mess with requrimenets
14:29:30 <boris-42> amaretskiy: e.g. rally package it self won't work
14:29:36 <andreykurilin> ok, let's forget about splitting api and cli:)
14:29:41 <boris-42> andreykurilin: without rally-clinet and vise versa
14:29:42 <amaretskiy> good
14:30:46 <andreykurilin> boris-42 amaretskiy: I don't think that spec will help match. It is easier to discuss stuff with implementation:)
14:30:57 <andreykurilin> *will help a lot
14:30:59 <amaretskiy> sure
14:31:42 <amaretskiy> so, spec or not to spec? :)
14:32:00 <ikhudoshyn> amaretskiy: not right now )
14:32:07 <amaretskiy> ok :)
14:32:10 <andreykurilin> +1 for ikhudoshyn:)
14:32:41 <ikhudoshyn> we'd review the patch and dicsuss cons if any, then maybe spec
14:32:41 <boris-42> andreykurilin: ok, so in any case try in mind to split rally and rally-cli in separated repo
14:32:51 <boris-42> andreykurilin: and understand how painful is that
14:32:52 <boris-42> =)
14:33:01 <andreykurilin> boris-42: I'll try;)
14:33:20 <amaretskiy> okay. I wait your reviews and being ready to start a spec :)
14:34:09 <andreykurilin> next topic?
14:34:28 <ikhudoshyn> we dont have any
14:34:34 <andreykurilin> I have one:)
14:34:40 <ikhudoshyn> ok)
14:34:49 <andreykurilin> so it is just an announcement
14:34:50 <ikhudoshyn> i actually do , too
14:35:02 <rvasilets_> We don't have topics more
14:35:16 <rvasilets_> #topic Open discussion
14:35:18 <andreykurilin> Last week a have a bit of free time and started writing release notes
14:35:31 <andreykurilin> I suppose we will have a release this week
14:35:56 <rvasilets_> )))
14:36:06 <andreykurilin> so let's concentrate on bug-fixes and move merge of huge changes to next week
14:36:26 <andreykurilin> it doesn't mean that we should ignore huge changes:)
14:36:31 <andreykurilin> just not merge them
14:36:35 <ikhudoshyn> there won't be any huge changes from me this week, i promise
14:36:42 <andreykurilin> :D
14:36:42 <ikhudoshyn> )
14:36:49 <andreykurilin> eom
14:36:59 <ikhudoshyn> one from me
14:37:02 <boris-42> andreykurilin: ok
14:37:06 <ikhudoshyn> a kind reminder stable and should be unchanged/versioned
14:37:12 <ikhudoshyn> oops
14:37:20 <ikhudoshyn> https://review.openstack.org/#/c/297020/
14:37:36 <ikhudoshyn> db refactoring is waiting for more eyes
14:37:48 <amaretskiy> I will review this soon
14:38:01 <ikhudoshyn> pls dont mind if its green - i'm working on migrations
14:38:18 <rvasilets_> Last week my eyes was tiered but I believe they are ready this week)
14:38:20 <ikhudoshyn> .. if its green or red
14:38:30 <ikhudoshyn> eom
14:38:32 <rvasilets_> or yellow)
14:38:41 <andreykurilin> ok, I'll try to fing time this week
14:39:24 <rvasilets_> amaretskiy, boris-42 what about to finish the meeting
14:39:29 <amaretskiy> +
14:39:31 <boris-42> +
14:39:36 <andreykurilin> boris-42: btw, my plans about release, do not mean that you can ignore patch about services...
14:39:36 <ikhudoshyn> +
14:39:41 <rvasilets_> Ok? thx to all
14:39:52 <rvasilets_> See you next time #endmeeting
14:40:01 <rvasilets_> #endmeeting