17:04:54 #startmeeting Rally 17:04:55 Meeting started Tue Feb 24 17:04:54 2015 UTC and is due to finish in 60 minutes. The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:04:56 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:04:58 The meeting name has been set to 'rally' 17:05:18 msdubov_: rvasilets__ ping 17:05:26 meteorfox: ping 17:05:39 boris-42:hi! 17:05:42 boris-42: hi 17:06:26 hi 17:07:16 let's wait for others 17:07:30 ok 17:08:44 amaretskiy: ping 17:08:48 hi 17:08:56 okay let's start 17:08:57 sorry for my being late 17:09:04 #topic Weekly updates 17:09:32 We removed ALL oclo-incubator code WOOT 17:09:47 -3K SLOC! 17:09:48 We add bunch of tests related to NovaKeyPairs 17:10:40 I fixed a quite nasty bug related to warnings (we shouldn't display any warnings in commands that prints ouput in special format json/...) 17:10:45 nice 17:11:00 also I saw, the *aas is out too? 17:11:29 was that part of the oclo-incubator code? 17:11:43 meteorfox: let's disccuss this a bit later 17:11:49 meteorfox: let me just finish overview=) 17:11:58 meteorfox: and then this will be first topic 17:12:05 alright sounds good 17:12:24 So we add new script for Rally gates it's written in python and supprots tags 17:12:34 Which will allow to use one job with multple tasks 17:13:01 that are testing keystone v3 / with existing users / various version of python 17:13:03 and so on 17:13:32 As well we set jobs for Mirantis OpenStack 17:13:47 and now we are testing every patch against *real* deployment 17:13:52 multinode=) 17:14:08 so we will be able to test VM migration stuff 17:14:09 in gates 17:14:42 So we are going to cut version 0.0.2 of Rally during this week 17:14:52 Okay let's discuss in details 17:14:59 #topic Rally AAS stuff 17:15:02 meteorfox: hey 17:15:19 boris-42: hey 17:15:23 meteorfox: so that AAS code was non working 17:15:28 meteorfox: it was written in pecan 17:15:35 meteorfox: and used oslo-incubator 17:15:40 we decided to chose another direction 17:15:44 use flask 17:15:53 boris-42: ah I see 17:15:55 and make AAS on top of rally.api 17:16:15 like a smart* code that adds 1-1 aas on top of rally.api 17:16:22 msdubov_: refactored it a bit 17:16:40 meteorfox: https://github.com/stackforge/rally/blob/master/rally/api.py so we will continue this work 17:16:41 awesome 17:17:15 meteorfox: so the idea is to avoid duplication of code 17:17:43 boris-42: awesome, this looks good. I'm exited for this feature 17:17:45 meteorfox: as well to reuse rally.cmd for bot (rally as cli tool and rally as a service) 17:17:52 meteorfox: you can help and work on it) 17:18:13 boris-42: :) sure 17:19:05 so let's to move to next topics 17:19:14 #topic new constant runner 17:19:20 msdubov_: any news ? 17:19:34 boris-42 yes 17:19:57 I believe I have written all the unit tests I wanted for the patch: https://review.openstack.org/#/c/155225/ 17:20:04 And it's no longer a WIP-patch 17:20:22 So boris-42 amaretskiy redixin You're welcome to review & test it 17:20:28 ok 17:20:49 I have also removed a functional test that checked that the number of iterations always grows 17:21:15 Because actually with the constant runner it may be the case that the iteration No. (i+1) starts before iteration No. i 17:21:30 (Not actually starts, but is logged) 17:21:49 So the iteration numbers in logs are not always strictly increasing 17:22:03 msdubov_: yep they are not 17:22:45 boris-42: it seems though that for the RPS runner they are often in-order :) 17:22:59 boris-42:Just an observation 17:23:11 boris-42: Basically that's it for the constant runner 17:23:22 the constant for duration runner, as discussed, will be modified later 17:23:37 msdubov_: ok 17:23:43 msdubov_: I will try to review it 17:23:48 #topic Murano patches 17:23:51 rvasilets__: hey hey 17:24:03 Hi 17:25:36 How i have separating from Murano base another patch its about refactor ImageGenarator and Validation 17:26:13 Just Murano stuff is okey now and its only wait to be separated from ImageGenerator 17:26:29 eom 17:27:04 rvasilets__: okay I hope you will separate it soon 17:27:15 I hope too 17:27:25 rvasilets__: it acutally doens't require any code changes 17:27:37 rvasilets__: just git 80lvl skills 17:27:37 =) 17:27:40 yes, just copy paste 17:27:48 rvasilets__: no there is no need to copy paste =( 17:27:49 and delete) 17:27:52 rvasilets__: nope.. 17:28:06 rvasilets__: I will tell you how to do it after meeting 17:28:20 okay let's move to next topic 17:28:25 #topic Scalable graphs 17:28:35 amaretskiy: how is your patch? realted to graphs? 17:28:45 last week i removed WIP from https://review.openstack.org/#/c/146814/ 17:28:50 this patch is working 17:29:06 however today it appeared in merge conflict 17:29:32 so I'm going to fix conflict + fix some tests (reviewed) tomorrow 17:29:43 so reviews are welcome 17:29:51 also there are >1k loc 17:30:00 so maybe i should split thi spatch 17:30:05 into 2 parts? 17:30:12 what do you think? 17:31:32 amaretskiy: I think it will be great 17:31:37 amaretskiy: to split it to 2 parts 17:31:37 okay 17:31:47 agreed 17:31:49 eom 17:32:05 amaretskiy: maybe as first step is deleting chart/histogram 17:32:18 amaretskiy: and moving it to charts (with refactoring) + tests 17:32:44 this is not a simple question 17:32:45 amaretskiy: it iwll help a lot 17:32:57 amaretskiy: why it is not simple question? it will be simple to do 17:33:15 i think sthis split should not take a lot of time, but i do not see if this possible 17:33:37 i will consider if it is possible to split thi spatch with minimal work 17:34:03 maybe it is a good idea to just add new charts classes + tests 17:34:20 without touching plot.py at all 17:35:01 since charts/histogram.py is completely rewritten 17:35:30 amaretskiy: yep you can make first charts 17:35:31 so i guess the first patch will simply add new charts classes 17:35:34 amaretskiy: with tests 17:35:39 yes 17:35:46 amaretskiy: and the remove old processing stuff and switch to new charts 17:35:50 amaretskiy: yep that makes sense 17:35:52 sure 17:36:13 amaretskiy: okay let' move to next stuff 17:36:25 #topic Rally Plugins 17:36:35 Okay I am working on unification of plugins 17:36:44 Finally I start seeing full picture 17:36:55 so I am going soon to push new patch 17:37:01 that changes plugin base 17:37:08 and allow to make from functions plugins 17:37:24 so this thing was the very last blocking thing that I didn't know how to resolve 17:37:31 as well I won't touch bese.context() 17:37:46 and instead of removing name I will just call plugin decorator inside it 17:38:06 so at least scenarios, context will be backward compbale 17:38:37 SLA stuff have to be refactored one more time 17:38:47 as well runners, deploy engines and server provides 17:39:08 but as far as I know nobody is writing those plugins so it's not a big deal to change them 17:39:34 After we switch to plugin base I will work on moving validation mechanism to plugin Base 17:39:49 this will allow us to use the same validation mechanism for all plugins that we have 17:40:06 boris-42, What will be the changes to SLA? 17:40:30 msdubov_: I will introduce @sla decorator 17:40:45 msdubov_: and remove option_name 17:41:29 boris-42: okay, thanks 17:43:00 so okay 17:43:06 #topic Open Discussion 17:43:13 anybody wants to discuss anything? 17:43:24 no 17:43:33 no 17:43:33 nope 17:44:22 no 17:45:25 Okay I don't have as well =) 17:45:29 Have a nice day=) 17:45:34 #endmeeting