00:05:20 <thinrichs> #startmeeting CongressTeamMeeting 00:05:21 <openstack> Meeting started Thu Sep 1 00:05:20 2016 UTC and is due to finish in 60 minutes. The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot. 00:05:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 00:05:24 <openstack> The meeting name has been set to 'congressteammeeting' 00:05:38 <aimeeu> hello everyone 00:05:43 <masahito> hi 00:05:44 <thinrichs> Agenda today… 00:05:49 <thinrichs> 1. bug in client? 00:05:55 <thinrichs> 2. newton3 00:05:57 <thinrichs> 3. status 00:06:01 <thinrichs> Anything else? 00:07:14 <thinrichs> #topic bug in client 00:07:22 * thinrichs trying to find link 00:07:43 <aimeeu> #link https://bugs.launchpad.net/congress/+bug/1619065 00:07:43 <openstack> Launchpad bug 1619065 in congress "Failure to push data through CLI" [High,New] 00:08:03 <aimeeu> thinrichs: still had it open from earlier 00:08:09 <thinrichs> There it is. thanks aimeeu 00:08:21 <thinrichs> I ran across this when getting ready to release the new client. 00:08:33 <thinrichs> The new feature addition is that we can push data through our CLI to push drivers. 00:08:39 <thinrichs> But I couldn't seem to get it ot work. 00:08:55 <thinrichs> masahito: am I creating the datasource correctly and pushing data correctly? 00:09:25 <thinrichs> (This is pressing since we're supposed to finalize the client by tomorrow and do a release.) 00:10:14 <masahito> I think the data schema of the request is not correct for doctor driver. 00:11:15 <masahito> #link https://github.com/openstack/congress/blob/master/congress/datasources/doctor_driver.py#L47 00:11:53 <masahito> this is the request body doctor's event table expects. 00:12:10 <thinrichs> Ok. Thanks. I'll fix up the schema, and assuming that works, release the client tomorrow. 00:12:30 <thinrichs> Glad it's likely to be user error. :) 00:12:38 <thinrichs> #topic newton3 00:12:56 <thinrichs> I haven't released newton3 yet. I'll do that tomorrow. 00:13:20 <thinrichs> Ideally we'd have all the new 'features' merged before newton3 00:13:30 <thinrichs> and only do bug fixing after newton3 00:14:19 <ekcs> great! 00:14:30 <ekcs> this can be considered either feature or fix: https://review.openstack.org/#/c/356806/ 00:15:01 <ekcs> This would be great to get in: #link https://review.openstack.org/#/c/354629/ 00:15:41 <thinrichs> Agreed that both those would be good to get into Newton. 00:16:12 <thinrichs> There seem to be a number of outstanding reviews that are mostly ready… 00:16:13 <thinrichs> https://review.openstack.org/#/q/project:openstack/congress+status:open 00:18:15 <thinrichs> So let's try to get the ones that are ready merged soon so that we can turn our attention to testing, bug fixing, and docs 00:18:52 <thinrichs> Let's say end of the week to get non-bugfixes in. 00:20:01 <thinrichs> That means we all are actively reviewing each others' code 00:20:34 <thinrichs> and thinking critically about whether a comment/improvement should stop the inclusion of the patch in Newton 00:20:46 <thinrichs> Any questions about newton3? 00:21:17 <ekcs> got it. 00:21:31 <thinrichs> #topic status 00:21:39 <thinrichs> ramineni_: want to start? 00:21:46 <ramineni_> sure 00:22:52 <ramineni_> I have adressed the comments on all of my open patches, but stuck in scary merge conflict :( will fix them up today 00:23:33 <ramineni_> Added a patch, to start mltiple PE and run existing tempest tests , but it seems to fail 00:23:47 <thinrichs> I'm intrigued: why is the merge conflict scary? 00:23:55 <ramineni_> have to check if its sync timing issue or any other issue 00:24:34 <ramineni_> thinrichs: need to again change multiple files to resolve them 00:25:23 <thinrichs> ramineni_: let's get that merged soon then. Those merge conflicts eat a bunch of time if your patch is touching many files. Much easier for other N patches to solve 1 merge conflict than for you to deal with N. 00:25:24 <ramineni_> thats it from my side 00:25:53 <ramineni_> ya, will do it today 00:25:56 <thinrichs> ramineni_: the tests for multiple PEs is super-crucial for this release 00:26:29 <ramineni_> yes, ill have a look today 00:26:29 <thinrichs> we need to get the HAHT working in at least a couple interesting scenarios 00:26:55 <thinrichs> We can all help you get that patch merged. Which one is it? 00:27:27 <ramineni_> https://review.openstack.org/#/c/362761/ 00:27:53 <ramineni_> by default i made it to true, to check if it works before adding a job in infra 00:28:51 <ramineni_> yesterday, couldnt find tim to debug it 00:29:07 <thinrichs> Sorry—I meant the patch that's got all the merge conflicts. 00:29:32 <ramineni_> oh.. 00:29:42 <ekcs> I’ll take a look at the tempest fails. It’s expected that some will fail from timing issues and stuff. 00:30:07 <ekcs> passing 20 out of 23 is actually very encouraging =) 00:30:19 <ramineni_> ekcs: right :) 00:30:21 <ekcs> great work. 00:30:37 <thinrichs> ramineni_: after a quick look, there might be some weird functionality that murano is expecting—like being able to insert a policy 00:30:49 <ramineni_> thinrichs: thats fine.. ill get it done today, they are series of dependent patches on each other , so its better ill do it 00:31:17 <thinrichs> ramineni_: okay; leave a comment if you want someone to get it merged sooner rather than later b/c it's causing so many conflicts 00:31:26 <ramineni_> sure, thanks 00:32:29 <thinrichs> ramineni_: ping me (irc and/or email) if you want my help with the murano failure 00:32:43 <thinrichs> masahito: want to give a status update? 00:32:49 <masahito> sure 00:33:01 <ramineni_> ok, sure 00:33:12 <masahito> 1. the patch of push driver support in client is merged 00:33:42 <masahito> 2. I'm working lazy datasource support in all OpenStack driver. 00:33:57 <masahito> tha't from my side 00:34:35 <masahito> s/tha't/that's/ 00:35:54 <thinrichs> Sounds good. 00:36:20 <thinrichs> aimeeu: what's new with you? 00:36:51 <aimeeu> Well not much. ;) Thanks everyone for the feed back on the HA Guide - it's been merged 00:37:50 <aimeeu> I've been reviewing code and trying to get PyCharm & DevStack configured for remote debugging so I can tackle the Congress dashboard bug 00:38:19 <aimeeu> that's it for me 00:38:41 <thinrichs> Thanks. Glad you got the HA guide merged. 00:39:02 <aimeeu> OH I did noticed that the keystone datasource driver is throwing an error 00:39:04 <thinrichs> I'm interested if you find a good way to debug devstack. I'm usually just using a text editor and the command line. 00:39:48 <thinrichs> ekcs: want to go next? 00:39:48 <aimeeu> If I don't figure it out tomorrow morning, I'll reach out to a resource at work (the guy who is doing the hands on in Barcelona) 00:40:03 <ekcs> I think this is ready to merge: #link https://review.openstack.org/#/c/356806/ 00:40:04 <ekcs> Not sufficient for detecting dropped update from an infrequently pushed DSD, but that can be addressed in future change. 00:40:15 <ekcs> On the integrated HA testing in test_congress, I've been slowed by many little issues expected when using an unorthodox approach, but I think mostly resolved. 00:40:15 <ekcs> Latest issue I am dealing with is that somehow I can't get congress to take the bind_port option from config file instead of the default port. Still investigating whether it's user error or bug. 00:40:24 <ekcs> That’s all. 00:41:47 <ekcs> I’m guessing testing/user error because ha_test does it fine on tempest. 00:42:34 <thinrichs> That patch of yours sounds like its ready to go into newton3 00:42:43 <thinrichs> One minute and I'll be back 00:43:19 <thinrichs> Back 00:43:59 <ekcs> yup I think so. 00:44:02 <thinrichs> Seems like we're all on a good trajectory to finish up major items for newton3 00:44:08 <thinrichs> So that's great! 00:44:27 <thinrichs> That's all I have for today, so we'll open it up for discussion. 00:44:30 <thinrichs> #topic open discussion 00:45:08 <ekcs> not a high priority, 00:45:18 <ekcs> but anyone know why the testr_results.html.gz stopped showing up in tempest tests? 00:46:11 <ekcs> It was there in this patch: https://review.openstack.org/#/c/358388/ 00:46:27 <ekcs> and then in the very next merged patch, it’s gone. https://review.openstack.org/#/c/356157/ 00:48:39 <thinrichs> Which gate job? I saw testr_results for py35 (I think) 00:49:24 <ekcs> in the tempest tests. like dsvm and murano. 00:50:48 <thinrichs> No idea. 00:50:52 <thinrichs> Maybe ping #infra 00:51:05 <thinrichs> We should check it's still not showing up. 00:51:21 <thinrichs> Maybe there's some setting that changed in how infra runs jobs 00:51:47 <ekcs> got it. 00:52:23 <ekcs> If there is nothing else: 00:52:28 <ekcs> does this config file look right? 00:52:29 <ekcs> [default] 00:52:30 <ekcs> bind_port = 4000 00:52:31 <ekcs> auth_strategy = noauth 00:54:54 <thinrichs> Looking at my default /etc/congress/congress.conf 00:55:06 <thinrichs> I see the auth_strategy = <blah> line 00:55:11 <thinrichs> I don't see the bind_port line 00:56:00 <ekcs> yea… I’m trying to override the default port for a second PE node. 00:56:30 <ekcs> but somehow congress still using old port. Don’t worry about it I’m pretty sure I’m doing something silly that’s causing it. 00:57:16 <thinrichs> Okay. 00:57:42 <thinrichs> Let us know if we can help. 00:58:42 <thinrichs> Seems that's the end of our time this week. 00:58:43 <thinrichs> Thanks all! 00:58:45 <thinrichs> #endmeeting