14:00:11 #startmeeting Rally 14:00:12 Meeting started Mon Sep 28 14:00:11 2015 UTC and is due to finish in 60 minutes. The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:15 The meeting name has been set to 'rally' 14:00:17 o/ 14:00:36 ping rvasilets_ ikhudoshyn e0ne 14:00:42 ping redixin 14:00:42 \o 14:00:44 hi 14:00:57 annyong haseyo 14:01:46 redixin: hey hey 14:02:50 okay let's start 14:03:16 #topic [redixin] rally-ci updates 14:03:47 hi 14:04:04 some updates was accumulated in rally-ci 14:04:19 now it supports cloud images 14:04:19 hi all 14:04:33 hi 14:04:45 also it support btrfs (previously was only zfs) 14:05:48 also added lots of small changes like timeouts and triggering jobs by ref-updated 14:06:22 current rally-ci configuration may be found here: https://github.com/redixin/rci-config/blob/master/config.yaml 14:06:26 redixin: so can you cut releases and make release notes 14:06:46 redixin: or somehow else peridoically wirte updates 14:07:03 redixin: just to keep process transparent for others? 14:07:06 it is not actually release. status of rally-ci may be seen here https://blueprints.launchpad.net/rally-ci 14:07:37 so there is some work to be done before first release 14:08:24 here is the main repository: https://github.com/redixin/rally-ci/ 14:09:31 so feel free to contact me if you want to contribute 14:09:59 redixin: ok 14:10:04 redixin: thanks for update 14:10:10 any questions guys? 14:10:30 no 14:10:45 no 14:12:08 #topic [boris-42] Rally 0.1.0 updates 14:12:20 Okay so finally we get Rally 0.1.0 released 14:12:28 we did a great job guys =) 14:12:30 woot 14:12:50 so we have fixed almost all non scalable code in Rally 14:13:00 and finished rally task abort command 14:13:23 improved a lot framework 14:13:29 and here are release notes https://rally.readthedocs.org/en/latest/release_notes/latest.html 14:13:50 I hope next release won't be delayed so much as this one=) 14:14:14 #topic [boris-42] About Rally 0.1.1 14:14:34 So I believe this week we should cut new Release 0.1.1 14:14:43 mostly because of serious bug in Rally reports 14:15:04 I would like to include murano benchmarks in this release 14:15:19 + patches releated to Tempest.conf generation 14:15:31 so it will be quite nice result for 1 week of development 14:15:32 -) 14:15:36 any questions? 14:15:42 any suggestions? 14:16:34 redixin: rvasilets_ redixin temujin_ ^ 14:16:36 should Task refactoring enter in 0.1.1 14:16:37 ? 14:16:37 andreykurilin__: ^ 14:16:42 include fuel-nodes scenario too 14:17:00 temujin_: so I believe we can include fuel-nodes 14:17:10 if we talk only about fixing task reports then its ok to cut realese in a week 14:17:26 ikhudoshyn: I believe it is not ready yet, hovewer I will start reviewing it 14:17:36 gotit 14:17:40 rvasilets_: why otherwise is not ok ? 14:18:02 boris-42, murano benchmarks was already finished then they too ok cut in a week 14:18:17 I think its possible to cut in a weeks 14:18:23 *week 14:19:29 so okay 14:19:46 let's move to next topic 14:20:38 #topic [boris-42] Critical bugs in reports 14:20:52 so rvasilets_ did you finish patch that I started? 14:21:28 Now I just want to add several more tests to MainStatsTableTestCase other ok 14:21:39 it wiil take not a lot of time now 14:22:01 I have almost finished this patrt 14:22:12 other is now on review 14:22:22 rvasilets_: great thanks for helping on this 14:22:47 guys please review this patch https://review.openstack.org/#/c/228403/ 14:22:54 it's qutei important 14:23:09 redixin: andreykurilin__ ikhudoshyn temujin_ ^ 14:23:16 will do 14:23:20 okay moving to next topic 14:23:23 already there 14:23:38 #topic [ylobankov] Creating Tempest resources 14:23:41 ylobankov: hi 14:23:48 hi guys 14:24:16 currently i am working on `rally verify` command 14:24:59 I have started work on this https://review.openstack.org/#/c/226652/ . Of course patch is in progress 14:25:45 ylobankov: so do you have any questions related to that patch? 14:25:57 I would like to discuss some issues. The first one is creating tempest resources 14:26:32 any ideas how to do it. Context or something like that? 14:27:29 ylobankov: so why not just creating before run and delete after run? 14:27:59 ylobankov, which resources do you mean? 14:28:10 boris-42: because we support cli which just generate config file, not launch verification 14:29:02 redixin: I suppose tempest specific roles, images and etc 14:29:12 boris-42: we have `rally verify genconfig` command. What should we do when we execute this one? redixin: I mean keystone roles, flavors and images 14:30:10 ylobankov, resources created by rally before (and for) tempest? or resources left by tempest? 14:30:22 Maybe, we can implement some context manager, which will check existence of resources from config file and create/remove them 14:30:35 redixin: for tempest launch 14:31:58 andreykurilin__: create/remove before each run `rally verify`? 14:32:25 ylobankov: yep 14:32:32 ylobankov: yep 14:32:55 ylobankov: we can as well store in rally.db reference to created resources so we will be able to cleanup them at any moment 14:33:09 or we can ask users "do you want to left tempest pre-created resources or not" 14:33:10 ylobankov: but it is not safe (because you can delete VM with rally for example) 14:35:26 It looks like we cannot create resources before each run `rally verify start` because we have `rally verify genconfig` 14:36:10 ylovankov: what about checking resources from config file before each run? 14:36:44 andreykurilin__: sounds resonable 14:38:22 it looks like at this point I don't have any questions. Any questions for me? 14:38:58 ylobankov: thanks for work on rally verify =) 14:39:34 boris-42: not at all :) 14:40:58 okay let's move to next topic 14:41:17 #topic [rvasilets] A small refactor of Rally install script 14:42:04 Okey. A week ago I get a task to remove psycopg2 from requrements 14:42:08 https://review.openstack.org/#/c/190931/ 14:43:23 But after delete from #link https://review.openstack.org/#/c/190931/4/requirements.txt requirenments 14:43:48 I found that there are also some packages that could be installed optionally too 14:44:25 All of them here #link https://github.com/openstack/rally/blob/master/install_rally.sh#L289 14:45:04 libpq-dev postgres-devel and postgressql93-devel 14:45:40 And I decide that we need to add manually there instalation here #link https://github.com/openstack/rally/blob/master/install_rally.sh#L400 14:45:55 But it takes a lot of lines more then 100 14:46:23 so I suggest to refactor script and add such function as install_pkg_manually 14:46:52 The changes that helped me with this are here #link https://review.openstack.org/#/c/190931/4/install_rally.sh 14:46:57 redixin: ^ could you take as well on this patch 14:47:29 Please say you thoughts about this refactor 14:47:32 eom 14:47:52 rvasilets_: okay my overall opinion is to try to find the ways to make that script shorter 14:47:55 rvasilets_: not longer=) 14:48:40 also there is exist paramater like easier 14:49:05 and not always shorter is easier) 14:50:50 rvasilets_: ok 14:50:52 Also this question is not intended to raise a debate on the needed of the script 14:51:02 it is out of scope 14:51:14 rvasilets_: I am not talking that we don't need script 14:51:21 rvasilets_: I am talking about it's complexity 14:51:40 hovever I will re-review it 14:51:48 okay seems like we dicussed everything 14:51:49 I have heard how guys are talkiong) 14:51:58 rvasilets_: ?) 14:52:09 this is more for them was said 14:52:49 andreykurilin__, redixin and some others don't like bash and want to rewrite it into python) 14:53:19 that was told to me when I have raised the same question to them) 14:54:02 the main problem of this script: rally team doesn't have a lot of bash-gurus 14:54:16 andreykurilin__: imho rewriting it to python will be bad idea 14:54:30 andreykurilin__: chris knows it well=) 14:54:41 This is a good reason to became a guru) 14:54:55 andreykurilin__: python imho is not a good language to install packages 14:55:27 Alexander said that he has 5000 line script experience on Bash) 14:55:38 *had 14:56:11 boris-42: after looking at our install script, I think that bash is not a good language to install packages too. 14:56:41 so okay andreykurilin__ lol 14:56:54 okay so we have to finish our meeting 14:57:01 we can continue disucssion in rally chat 14:57:05 ok 14:57:06 ) 14:57:07 see you guys 14:57:11 #endmeeting