14:01:05 <amaretskiy> #startmeeting Rally
14:01:06 <openstack> Meeting started Mon Sep 21 14:01:05 2015 UTC and is due to finish in 60 minutes.  The chair is amaretskiy. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:09 <openstack> The meeting name has been set to 'rally'
14:01:16 <redixin> sup
14:01:18 <amaretskiy> hi all
14:02:24 <rvasilets> o/
14:02:35 <ikhudoshyn> o/
14:03:46 <amaretskiy> let's start
14:04:04 <amaretskiy> #topic New Rally release model
14:04:38 <amaretskiy> I believe all have seen email from boris-42 regarding new release model
14:04:44 <andreykurilin> hi
14:05:12 <amaretskiy> now we have flexible ability to publish major and minor releases
14:05:47 <amaretskiy> does anyone have comments about that?
14:05:57 <redixin> nope
14:07:26 <amaretskiy> okay
14:08:04 <yfried_> hi
14:08:37 <amaretskiy> hi
14:10:03 <amaretskiy> so we have rule for major releases (N).(N).(n), minor releases (n).(n).(N), bug fixes and releases for demand - (n).(n).(n)-(N).(N)
14:10:20 <amaretskiy> i believe I'm not wrong with this short description :)
14:10:37 <amaretskiy> okay, let's proceed
14:10:51 <amaretskiy> #topic Discuss Rally Tags
14:11:17 <amaretskiy> We have a proposal from boris-42 for 3 tags
14:11:31 <amaretskiy> they are: covered-by-rally, has-rally-gates, certified-by-rally
14:11:45 <amaretskiy> I think we should be ready to discuss that with Boris
14:13:06 <amaretskiy> for example, first tag "covered-by-rally" is not clear for me - if we set this tag for some project, this definitely does not guarantee that all API are covered by rally
14:13:20 <amaretskiy> does anybody have ideas about that?
14:13:23 <kbaikov> o/
14:14:13 <rvasilets> But covered all Api that is in Rally)
14:14:36 <e0ne> hi!
14:14:40 <amaretskiy> hi
14:15:12 <amaretskiy> andreykurilin, redixin - what do you think? ^
14:15:23 <rvasilets> I thin that the best is the first idea. "works in rally"
14:15:48 <rvasilets> Its smart and means exactly what is in tag
14:15:50 <amaretskiy> actually it seems that Boris has kicked this tag
14:16:53 <andreykurilin> rvasilets: "works in rally" doesn't say load in which it works
14:17:38 <amaretskiy> okay, we have a proposal from Boris regarding 3 tags
14:17:48 <amaretskiy> let's be ready to discuss that
14:17:50 <rvasilets> with two words I don;t think that we can describe the exact load
14:18:10 <amaretskiy> also read his email regarding tags
14:18:27 <amaretskiy> [openstack-dev] [openstack-operators][tc][tags] Rally tags
14:18:44 <rvasilets> but we can think like: rally = means high load, means that it would work in any case
14:19:30 <amaretskiy> rvasilets: I think we do not mean high load, especially on gates based on devstack
14:20:57 <amaretskiy> rvasilets: also I've submitted a patch to keystone gates, and there was no high load since I was asked to make rally job running about 20 mins
14:21:01 <amaretskiy> okay, does anybody have something to add about tags?
14:21:27 <rvasilets> ok. Rally its a competition for professionls. What if "ride in rally"?)
14:22:12 <amaretskiy> rvasilets: okay, let's discuss that later with Boris
14:22:24 <rvasilets> ok
14:22:24 <amaretskiy> okay, proceed to next topic
14:22:41 <amaretskiy> #topic Status for implementation of 'rally task abort' command
14:23:02 <amaretskiy> andreykurilin: what is the status for this patch?
14:23:23 <andreykurilin> py27: commands succeeded
14:23:24 <andreykurilin> congratulations :)
14:23:35 <andreykurilin> just need to publish new patch-set
14:23:36 <amaretskiy> :)
14:23:45 <andreykurilin> and check pep8
14:23:51 <amaretskiy> great
14:24:48 <andreykurilin> new patch-set is ready
14:24:59 <rvasilets> how about tox -e cli?)
14:25:18 <andreykurilin> rvasilets: I don't launch it locally:)
14:25:24 <rvasilets> I beliave that we will cut release after merging this)
14:25:27 <amaretskiy> cool. we need this patch since it is a release-blocker
14:25:34 <amaretskiy> okay, proceed
14:25:49 <andreykurilin> yes, it looks like it last release-blocker
14:27:22 <amaretskiy> #topic member role in users context - difference for v2 and v3 api
14:27:22 <amaretskiy> stpierre hi
14:27:29 <stpierre> hey
14:27:48 <stpierre> so boris-42 kind of weighed in on this, and it sounds like he wants to go with the v2 behavior -- that is, no default group
14:27:54 <stpierre> let me find the code review real quick
14:28:06 <stpierre> https://review.openstack.org/#/c/220441/
14:28:13 <stpierre> #link https://review.openstack.org/#/c/220441/
14:28:34 <stpierre> anyone have any problems with rolling back the keystone v3 wrapper behavior that automatically adds users to the 'member' role?
14:29:41 <amaretskiy> stpierre: okay. personally I use v2 on my local cloud :(
14:29:47 <stpierre> yep, me too
14:31:54 <stpierre> hearing no opposition, i guess we follow our BDFL's lead and get rid of the v3 wrapper behavior
14:32:25 <amaretskiy> stpierre: does keystone team have proper bug reported?
14:32:45 <stpierre> the only bug report i'm aware of is this one:
14:32:48 <stpierre> #link https://bugs.launchpad.net/rally/+bug/1492177
14:32:49 <openstack> Launchpad bug 1492177 in Rally "Users created without the member role with keystone API V2" [Undecided,Invalid]
14:33:27 <andreykurilin> gg, I marked this bug as Invalid:)
14:33:58 <stpierre> perfect :)
14:34:33 <amaretskiy> it seems that we should report proper bug to keystone, shouldn't we?
14:34:46 <stpierre> what bug exactly?
14:35:06 <amaretskiy> afaik keyston v3 api works wrong... ?
14:36:02 <stpierre> oh, is that what this is trying to work around?
14:36:34 <amaretskiy> stpierre: i do not know
14:36:36 <stpierre> i just assumed someone didn't want to add the member role using the roles context :)
14:36:56 <stpierre> i guess i can bring up a devstack and TIAS easily enough
14:37:17 <amaretskiy> stpierre: okay, waiting updates from you. thanks!
14:37:26 <andreykurilin> stpierre: maybe we should merge users and roles contexts ?
14:38:36 <stpierre> maybe -- that would certainly be a more elegant way to handle creating a default role
14:39:51 <amaretskiy> merge users and roles contexts - this looks like huge change and should be discussed with boris-42
14:39:59 <stpierre> oh very yes
14:40:12 <andreykurilin> also, I don't know how it should work for existing users
14:40:44 <stpierre> i think for now it'd be good if we could make the behavior consistent across the keystone wrapper; anything bigger than that is out of scope for the moment
14:41:02 <amaretskiy> great
14:41:51 <amaretskiy> okay, let's proceed
14:41:58 <amaretskiy> #topic [Swift] Add objects context class
14:42:08 <amaretskiy> #link https://review.openstack.org/#/c/179578/
14:42:32 <amaretskiy> according to our gerrit dashboard, this patch is important for release
14:42:43 <amaretskiy> does anybody know status for it?
14:43:07 <andreykurilin> amaretskiy: gerrit dashboard is outdated
14:43:16 <amaretskiy> andreykurilin: okay
14:43:28 <amaretskiy> i guess we can proceed :)
14:43:36 <amaretskiy> #topic Free discussion
14:43:46 <andreykurilin> #topic i18n
14:44:17 <andreykurilin> As you know, Rally uses oslo.i18n library
14:44:41 <andreykurilin> also, all user messages are wrapped by _(), _LE()...
14:44:43 <andreykurilin> BUT
14:44:56 <andreykurilin> But Rally doesn't have any pot file
14:45:02 <andreykurilin> we don't have any translations
14:45:57 <amaretskiy> andreykurilin: I think before starting writing .po files we should inspect all files and put translations for some messages
14:45:59 <andreykurilin> _() is required for any messages in rally, but it doen't work at all. just redundant wrappers fir now
14:46:38 <andreykurilin> amaretskiy: this process can be automated
14:46:41 <andreykurilin> and should be
14:46:59 <dims__> andreykurilin: https://wiki.openstack.org/wiki/Translations#Setup_Zanata_server.2C_import_and_export_of_translations
14:47:09 <amaretskiy> andreykurilin: need a hacking rule?
14:47:09 <andreykurilin> dims__: thanks:)
14:47:20 <andreykurilin> amaretskiy: see a link above
14:48:33 <andreykurilin> So we have 2 ways: finish configure of i18n or remove it at all for now
14:49:11 <amaretskiy> andreykurilin: first way looks better :)
14:49:46 <amaretskiy> but i do not see anything about hacking in the link above
14:50:10 <rvasilets> First would be easier)
14:51:33 <amaretskiy> I'm familiar with Django  - there is a makemassages command for scanning project and collecting all messageids
14:52:32 <amaretskiy> okay, this topic looks like required boris-42 attention as well
14:54:01 <amaretskiy> does anybody have something to add regarding i18n?
14:54:22 <rvasilets> nothing from me
14:55:31 <amaretskiy> okay
14:55:47 <amaretskiy> #topic Free discussion
14:56:34 <amaretskiy> any more comments or questions?
14:57:53 <amaretskiy> let's finish meeting :)
14:57:58 <amaretskiy> #endmeeting