14:01:00 <boris-42> #startmeeting Rally
14:01:00 <openstack> Meeting started Mon Jun 29 14:01:00 2015 UTC and is due to finish in 60 minutes.  The chair is boris-42. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:01 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:03 <openstack> The meeting name has been set to 'rally'
14:01:11 <boris-42> ping andreykurilin__ rvasilets_
14:01:55 <andreykurilin__> зщтп
14:01:58 <andreykurilin__> pong
14:01:59 <andreykurilin__> :)
14:02:15 <boris-42> andreykurilin__: you are on holidays no?
14:02:31 <andreykurilin__> boris-42: yes, but I'm online now:)
14:02:36 <boris-42> kun_huang: ping
14:02:43 <rakhmerov> hi
14:02:54 <kun_huang> :)
14:03:00 <boris-42> kun_huang: hi hi
14:03:05 <boris-42> stpierre: ping
14:03:07 <stpierre> pong
14:03:11 <boris-42> stpierre: hi there
14:03:16 <kun_huang> actually fixtures upgrades block us again...
14:03:32 <boris-42> kun_huang: hm?
14:03:46 <kun_huang> http://logs.openstack.org/32/193432/17/check/gate-rally-python26/5ec974d/console.html
14:03:46 <kun_huang> block unit test
14:04:23 <andreykurilin__> #link https://review.openstack.org/#/c/196623/
14:04:29 <kun_huang> there are some hidden bugs in rally
14:04:50 <boris-42> kun_huang: hm
14:04:51 <andreykurilin__> kun_huang: what kind of bugs?
14:04:53 <boris-42> kun_huang: https://review.openstack.org/#/c/196493/
14:04:59 <boris-42> kun_huang: my patch works well
14:05:06 <boris-42> kun_huang: except trivial bug in pep
14:05:38 <kun_huang> https://github.com/openstack/rally/blob/master/tests/unit/plugins/openstack/scenarios/glance/test_utils.py#L50
14:05:41 <kun_huang> this line
14:06:40 <boris-42> kun_huang: and?
14:06:48 <boris-42> kun_huang: it doesn't fail for me..
14:07:21 <kun_huang> going into your tox env, pip install fixtures --upgrade, back to normal env
14:07:30 <kun_huang> run tox -e py26, failed
14:07:43 <kun_huang> fixtures upgraded today...
14:07:59 <kun_huang> https://review.openstack.org/#/c/196086/
14:08:00 <boris-42> kun_huang: omh
14:08:02 <boris-42> kun_huang: omg
14:08:08 <kun_huang> the latest one, failed again
14:08:11 <boris-42> kun_huang: so many upgrades...
14:08:37 <kun_huang> the failed two test cases both have "self.get_fm.cleanUp()"
14:08:46 <kun_huang> I don't understand on this ...
14:09:50 <boris-42> kun_huang: that one seems like a bug
14:09:54 <boris-42> kun_huang: just need to remove that line
14:10:00 <boris-42> kun_huang: that you specified
14:10:18 <kun_huang> I'm not sure for that line
14:10:23 <kun_huang> are you sure for that?
14:11:45 <boris-42> kun_huang: nope
14:12:01 <boris-42> kun_huang: seems like kiran is working on it
14:12:07 * boris-42 kun_huang: https://bugs.launchpad.net/rally/+bug/1469644
14:12:07 <openstack> Launchpad bug 1469644 in Rally "Rally jenkins fails due to fixtures 1.3.0" [Undecided,In progress] - Assigned to Kiran (rkiran)
14:12:53 <andreykurilin__> boris-42: https://review.openstack.org/#/c/196623/
14:13:09 <andreykurilin__> boris-42: we can't merge such fix:(
14:13:10 <kun_huang> we two found it in my afternoon
14:13:10 <boris-42> andreykurilin__: that one is not a proper fix
14:13:30 <boris-42> kun_huang: ok I will investigate it today
14:13:36 <kun_huang> btw, removing that line can't fix this
14:13:46 <kun_huang> Traceback (most recent call last):
14:13:46 <kun_huang> File "tests/unit/plugins/openstack/scenarios/nova/test_utils.py", line 58, in test_failed_server_status
14:13:46 <kun_huang> server_manager.create("fails", "1", "2"))
14:13:46 <kun_huang> File "/Users/Gareth/celtics/rally/.tox/py27/lib/python2.7/site-packages/testtools/testcase.py", line 422, in assertRaises
14:13:46 <kun_huang> self.assertThat(our_callable, matcher)
14:13:46 <kun_huang> File "/Users/Gareth/celtics/rally/.tox/py27/lib/python2.7/site-packages/testtools/testcase.py", line 435, in assertThat
14:13:46 <kun_huang> raise mismatch_error
14:13:47 <kun_huang> testtools.matchers._impl.MismatchError: <MagicMock name='get_from_manager()' id='4466724112'> returned <MagicMock name='get_from_manager()()' id='4466814096'>
14:13:47 <kun_huang> Ran 1661 tests in 21.935s (-0.247s)
14:13:47 <boris-42> rakhmerov: thanks for fixing mistral client
14:13:48 <kun_huang> FAILED (id=46, failures=2)
14:13:48 <kun_huang> error: testr failed (1)
14:13:51 <boris-42> kun_huang: please
14:13:56 <boris-42> kun_huang: use paste
14:14:04 <kun_huang> okay :)
14:14:16 <rakhmerov> boris-42, np. My apologies again for the mess..
14:15:21 <boris-42> so okay we will fix this
14:15:26 <boris-42> let's move to next topics
14:15:58 <boris-42> #topic Import Optimization
14:16:20 <boris-42> I almost finished the patch that optimizes imports
14:16:32 <boris-42> https://review.openstack.org/#/c/196493/
14:17:13 <boris-42> so it just imports plugins if they are required
14:17:29 <boris-42> as a result I saw that it's not enough to fix the issue and it takes about 300ms to start Rally
14:17:40 <kun_huang> seems nice
14:17:44 <boris-42> another big issue is sqlalchemy that starts about 80ms
14:17:59 <boris-42> so I will try to hack more and reduce time at least to 200ms
14:18:14 <boris-42> and hope that that will be enough to do live bash completition
14:19:32 <boris-42> andreykurilin__: stpierre ^
14:20:36 <andreykurilin__> boris-42: "+58, -24" is quite small change for you:)
14:20:39 <stpierre> nice
14:21:02 <boris-42> andreykurilin__: LOL
14:21:06 <boris-42> andreykurilin__: I didn't finished it
14:21:12 <andreykurilin__> :D
14:21:18 <boris-42> andreykurilin__: I will need one more small change to +2500 -25000
14:21:20 <boris-42> =)
14:21:29 <boris-42> albertw: hi there
14:21:32 <albertw> hi
14:21:44 <boris-42> albertw: it's our regular weekly meeting
14:22:00 <boris-42> albertw: btw gates are against broken with something *new* released
14:22:10 <boris-42> albertw: so no worries about -1 from Jenkins =(
14:23:12 <albertw> boris-42: thanks. I was wondering why things were failing.
14:23:27 <boris-42> albertw: last two weeks was hell for us
14:23:39 <andreykurilin__> openstack is too stable for our gates :(
14:23:45 <boris-42> albertw: neutron 2 times, heat,  mistral, fixtures..
14:24:01 <andreykurilin__> boris-42: you worgot ceilometer
14:24:04 <andreykurilin__> forgot
14:24:05 <boris-42> andreykurilin__: ah yep
14:24:22 <boris-42> albertw: ^ so we can't do actually anything with this except fix/wait_for_fixes
14:26:07 <boris-42> so okay
14:26:37 <boris-42> #topic Open Disucssion
14:26:50 <boris-42> Actually, as msdubov left team
14:27:09 <boris-42> we should find somebody who is able to do Agenda for each meeting
14:27:49 <boris-42> stpierre: andreykurilin__ rvasilets_ ^
14:28:10 <boris-42> so everybody will be able to attend to meeting
14:28:41 <stpierre> this meeting is at the start of my day, so i vote for someone else so that i don't have to worry about it friday afternoon before leaving for the weekend.
14:28:51 <stpierre> if neither andrey or roman can do it, of course i'm happy to step in
14:29:08 <andreykurilin__> boris-42: do we have agenda at mailing-list ?
14:29:18 <andreykurilin__> *only at mailint-list
14:29:19 <boris-42> andreykurilin__: nope
14:29:59 <rvasilets_> if you show me how to do this
14:30:04 <rvasilets_> I'll try
14:30:23 <boris-42> rvasilets_: ok
14:30:30 <boris-42> rvasilets_: so it's simple https://wiki.openstack.org/wiki/Rally we should fix this page
14:30:48 <boris-42> rvasilets_: to add somewhere whole paragraph about meetings agenda
14:31:10 <boris-42> rvasilets_: then we should create a sepcial wiki page for that
14:31:51 <boris-42> rakhmerov: https://wiki.openstack.org/wiki/Meetings/Rally
14:31:53 <boris-42> rakhmerov: sorry
14:31:56 <boris-42> rvasilets_: ^
14:32:01 <boris-42> rvasilets_: there is already
14:32:11 <boris-42> rvasilets_: so the idea is to collect all topics that we should disucss
14:32:31 <boris-42> rvasilets_: regularly and write them there
14:32:35 <andreykurilin__> it's a bit outdated:D
14:32:45 <boris-42> andreykurilin__: yep
14:32:53 <boris-42> andreykurilin__: but we need to keep full history
14:33:00 <boris-42> andreykurilin__: with links to the logs
14:33:06 <andreykurilin__> oh...
14:33:36 <boris-42> andreykurilin__: yep
14:33:42 <boris-42> rvasilets_: are you ready ?)
14:33:45 <rvasilets_> you are talking about this page ?
14:33:50 <rvasilets_> https://wiki.openstack.org/wiki/Rally#Updates
14:34:37 <boris-42> rakhmerov: nope
14:34:46 <andreykurilin__> rvasilets_: https://wiki.openstack.org/wiki/Meetings/Rally
14:34:49 <boris-42> rvasilets_: nope
14:34:56 <boris-42> rakhmerov: sorry one more time
14:35:08 <rvasilets_> I think it would be better to swap holder of agenda sometimes. Because as I understand its boring)
14:36:16 <boris-42> rvasilets_: yep that sounds good
14:36:41 <boris-42> rvasilets_: e.g. everybody can do it 2 weeks
14:36:54 <boris-42> rvasilets_: and change with next
14:37:04 <rvasilets_> ok I'll start
14:37:05 <boris-42> andreykurilin__: stpierre kun_huang ^
14:37:27 <stpierre> ok, sounds good
14:37:34 <andreykurilin__> ok, sound reasonable
14:37:40 <rvasilets_> you will show me how to do it right in private messages)
14:38:23 <kun_huang> no problem :)
14:39:40 <boris-42> okay nice
14:39:55 <boris-42> so we will make a schedule and we will keep agends/logs up to date
14:40:03 <boris-42> and make this meeting more useful
14:40:29 <boris-42> do we have something to discuss?
14:41:08 <andreykurilin__> no :(
14:41:51 <kun_huang> boris-42: there is a set of patches about sahara/swift
14:41:56 <kun_huang> need review
14:41:56 <kun_huang> https://review.openstack.org/#/c/177732/
14:42:51 <andreykurilin__> boris-42: what about adding regular topic "stuck reviews"?
14:43:16 <boris-42> andreykurilin__: ya that sounds good
14:43:37 <boris-42> kun_huang: yep yep we have issues with review throughput...
14:43:52 <kairat_kushaev> andreykurilin__: How about heat issues to analyze=)
14:44:02 <boris-42> kun_huang: ?
14:44:08 <boris-42> kairat_kushaev: ? )
14:45:03 <kun_huang> I'm working on review work now
14:45:14 <kairat_kushaev> boris-42: I mean when can we proceed to analyze why heat is failing
14:45:26 <kun_huang> but I don't know sahara details
14:45:30 <andreykurilin__> kairat_kushaev: heh. today is a holiday in Ukraine. non-working day. :P
14:45:46 <kairat_kushaev> andreykurilin__: OK=)
14:45:46 <boris-42> kairat_kushaev: heh I really don't know why it started failing...
14:46:55 <kairat_kushaev> boris-42: We did some investigation but I need to understand what nova-network is doing at some moment. Anyway I will pink Andrey this week
14:47:09 <andreykurilin__> boris-42, kairat_kushaev: Maybe we can move heat job from nova-network to neutron?
14:47:23 <boris-42> andreykurilin__: yep that seems like a good idea
14:47:23 <kairat_kushaev> boris-42: I have a second question that was raised during review of some patches
14:47:30 <boris-42> kairat_kushaev: ?
14:47:31 <andreykurilin__> And I'll try to debug nova-network this week
14:47:40 <kairat_kushaev> andreykurilin__: seems reasonable
14:48:18 <kairat_kushaev> boris-42: OS is updating very fast. So what versions we should support in rally
14:48:33 <andreykurilin__> kairat_kushaev: all
14:48:34 <andreykurilin__> :(
14:48:43 <andreykurilin__> * as much as it possible
14:49:18 <boris-42> andreykurilin__: +2
14:49:20 <stpierre> can we reduce that to "all versions supported by upstream"?
14:49:28 <boris-42> stpierre: nope
14:49:33 <stpierre> i don't feel like we should particularly care if rally works against grizzly (it doesn't)
14:49:54 <boris-42> stpierre: so there is people that use grizzly and so on..
14:50:22 <andreykurilin__> and if you install rally in virtual env, it'll work with grizzly
14:51:38 <andreykurilin__> if we create stable branches(stable/icehouse, stable/juno and etc), we will add a lot of extra stuff(backporting patches)
14:51:51 <stpierre> i certainly agree that's a bad approach
14:52:05 <boris-42> stpierre: so let's try to imporve framework
14:52:21 <stpierre> but we don't have the CI infrastructure for making sure that rally works on older versions, and i think it's foolhardy to pretend that we can make our scenarios work for everything
14:52:22 <boris-42> stpierre: so we will be able to support old versions in more transparent way
14:52:42 <boris-42> stpierre: so better framework / better ci
14:52:47 <boris-42> stpierre: can help us
14:53:06 <stpierre> it can help us, but scope that expands without bound is not possible to satisfy.
14:53:20 <stpierre> i.e., if someone filed a bug against a scenario that doesn't work on Diablo, should we care?
14:53:22 <stpierre> (no.)
14:53:42 <boris-42> stpierre: if it is possible to fix, why not?)
14:54:16 <boris-42> stpierre: in any case by trying to support older version we are facing real production issues
14:54:30 <boris-42> stpierre: knowing such issues allows us to make better framework
14:54:42 <stpierre> a) we've got better things to do than set up diablo openstack to test the fix; b) we have no diablo CI resources, so we can't ensure that it stays fixed; c) that's going to *greatly* overcomplicate the code in the future because the APIs have changed significantly since older versions
14:54:59 <kairat_kushaev> +1
14:55:06 <stpierre> kairat_kushaev and i ran into an issue just last week where we've got probably two dozen extra lines of hacky code in one of the heat scenarios because kilo heat is good and juno heat is terrible
14:55:12 <stpierre> and that's *juno*
14:55:43 <stpierre> ("good" might be overstating it, but you get the idea)
14:55:48 <kairat_kushaev> =)
14:55:56 <taturn> hi
14:56:17 <taturn> anyone?
14:56:19 <andreykurilin__> taturn: hi
14:56:19 <taturn> so
14:56:23 <stpierre> i think we need to limit ourselves to what we can realistically test
14:56:29 <boris-42> taturn: hi
14:56:34 <taturn> cool
14:56:40 <taturn> i am new here
14:57:03 <boris-42> taturn: so welcome=)
14:57:07 <taturn> and the bad thing for me is i am not good at english
14:57:13 <taturn> thx
14:57:25 <taturn> by the way i am from china
14:57:56 <kairat_kushaev> Perhaps we need some minimal requirements
14:58:00 <stpierre> right now, that means four versions, which IMO is more than enough to keep us very busy
14:58:02 <boris-42> stpierre: agree with this
14:58:20 <boris-42> stpierre: in any case our framework sux
14:58:34 <andreykurilin__> stpierre: simple example. Mirantis just released mos-6.1, which is "kilo", so most of mirantis customers use juno today and will start upgrading to "liberty" in near future. I suppose there is a lot of such examples
14:58:40 <boris-42> stpierre: to have scenarios for specific versions..
14:58:59 <andreykurilin__> *will start upgrading to kilo
14:59:10 <boris-42> andreykurilin__: and we have customers that use grizlly
14:59:11 <boris-42> LOL
14:59:12 <stpierre> boris-42: yeah, that will help
14:59:17 <stpierre> most of our customers are on grizzly
14:59:28 <boris-42> stpierre: so we need to keep grizzly wokring
14:59:34 <stpierre> no, we really don't
14:59:45 <boris-42> stpierre: =)
14:59:53 <boris-42> so in any case
14:59:56 <boris-42> we should finish
14:59:59 <taturn> lol
15:00:00 <boris-42> this meeting
15:00:02 <stpierre> even if we just test what's currently supported, we're going to be testing four versions (and troubleshooting icehouse devstack issues, which is terrible enough)
15:00:14 <taturn> i donot know what u r saying
15:00:17 <taturn> lol
15:00:17 <boris-42> #endmeeting