16:00:47 #startmeeting oslo 16:00:48 Meeting started Mon Nov 21 16:00:47 2016 UTC and is due to finish in 60 minutes. The chair is harlowja_at_home. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:50 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:00:53 The meeting name has been set to 'oslo' 16:00:58 ihrachys> will do , thanks 16:01:06 ihrachys, Ok. 16:01:16 courtesy ping for amotoki, amrith, bknudson, bnemec, dansmith, dhellmann, dims 16:01:16 courtesy ping for dougwig, e0ne, flaper87, garyk, gcb, GheRivero, haypo 16:01:16 courtesy ping for HenryG, jd__, jecarey, johnsom, jungleboyj, kgiusti, kragniz 16:01:16 courtesy ping for lifeless, lintan, lxsli, Nakato, ozamiatin, rbradfor, redrobot 16:01:16 courtesy ping for rpodolyaka, sergmelikyan, sileht, spamaps, sreshetnyak, sreshetnyak, stevemar 16:01:18 courtesy ping for therve, thinrichs, toabctl, viktors, zhiyan, zzzeek 16:01:18 o/ 16:01:21 o/ 16:01:22 o/ 16:01:25 o/ 16:01:27 o/ 16:01:27 o/ 16:01:29 hi 16:01:36 hi hi 16:01:37 o/ 16:01:37 partially here o/ 16:01:40 o/ 16:01:43 hi hi hi 16:02:19 hi 16:02:50 #topic Red flags for/from liaisons 16:03:04 Nothing to report from lbaas 16:03:22 thx johnsom 16:03:37 nothing from keystone that I know of. 16:04:14 none from trove 16:04:16 and hello all 16:04:21 o/ 16:04:34 hello amrith :-P 16:04:49 \./ 16:05:36 alright, doesn't seem like any current/active issues, so that's good! 16:05:52 #topic Releases for Ocata 16:06:33 so i've been working through getting the periodic jobs back into shape before getting out a bunch of releases 16:06:48 #link http://status.openstack.org/openstack-health/#/?groupKey=build_name&resolutionKey=hour&searchProject=-with-oslo 16:07:07 * harlowja_at_home is hoping https://github.com/openstack-infra/project-config/commit/fe21f63cc2e26b78 fixes a bunch of issues there 16:07:30 lots of red there 16:08:00 ya, afaik mainly due to pulling down newer versions of non-oslo libraries that are busting things 16:08:18 needs upper-constraints 16:08:23 ya 16:08:51 the project-config change i did should now use the upper-constraints file 16:08:59 and then i hope get rid of http://logs.openstack.org/periodic/periodic-cinder-py35-with-oslo-master/ef2f5a8/console.html#_2016-11-20_06_08_59_144306 and similar 16:09:39 so if people are wondering why there hasn't been a major bulk oslo release just yet, this is why (trying to get those to be somewhat healthy first) :) 16:11:13 otherwise does anyone need a specific release of a library for ocata that i should know about? 16:11:29 https://review.openstack.org/#/c/398947/ (for newton for oslo.messaging) is one i know about 16:11:54 and https://review.openstack.org/#/c/398946/ 16:12:18 * rpodolyaka would like to see an oslo.db release, but it can wait before the jobs are green again 16:12:48 rpodolyaka, cool, sounds good 16:14:07 #topic Reviews needing review 16:14:25 any reviews people want some more eyes on? 16:14:34 #link https://review.openstack.org/#/c/398875/ is an interesting one 16:15:00 (but it's more of a governance discussion) 16:16:30 kgiusti, should people start looking at https://review.openstack.org/#/c/396738/ or should they wait a little bit :) 16:16:54 Yes - def take a look 16:17:07 it's not passing jenkins for an unrelated issue 16:17:40 but this is essentially a doc change - and a guess at how the driver API is supposed to work 16:17:44 :) 16:17:59 think of it as a 'straw-person' 16:18:03 lol 16:18:18 looking for folks to set it on fire.... err - correct and clarify it. 16:18:34 sileht, ^ set it on fire! 16:18:36 lol 16:19:53 #topic PTG 16:20:17 has anyone started making plans for going (or not going?) to the PTG in atlanta yet? 16:20:28 I am planning to attend 16:20:31 #link https://www.openstack.org/ptg/ 16:20:50 nothing booked yet but planning on being there.... 16:21:07 hopefully they don't run out of spots for attendees. 16:21:22 I'm hoping to be there but haven't gotten approval yet. 16:21:44 harlowja_at_home : I'll be there 16:21:55 ya, i gotta poke my manager, see about getting the travel stuff approved earlier rather than later :-P 16:22:24 monday - tuesday seems to be the oslo times/days 16:23:08 that makes sense 16:23:11 FYI, when you click the register/tickets link it tells you how many slots are left, there are 359 left now 16:23:22 I'll be splitting my time between oslo, release, and goals stuff those days 16:23:24 ah, nice, good to know :) 16:23:40 dhellmann, sounds busy, ha 16:24:02 yeah, it didn't really solve the "lots of overlapping interests" problem for me :-) 16:24:16 Yeah, same here 16:24:25 ya, i know what u mean :) 16:24:32 need cloning tech 16:24:41 we're still working on some sort of notification system for advertising when certain topics will be discussed in each room 16:25:05 SMS based? 16:25:15 That would be neat, though I can see my phone exploding... 16:25:18 lol 16:25:37 yeah, I think they were talking about an etherpad or something similar 16:27:11 cools 16:27:25 #topic Open discussion 16:27:41 amrith, thx again for helping out last week when i wasn't feeling to well 16:27:50 hello 16:27:52 amrith, is there anything needed in regard to the prior topic? 16:27:59 tovin07, hi 16:28:04 anyone works with OSProfiler? 16:28:20 hi harlowja_at_home sorry was on the phone 16:28:21 i haven't recently, i'm sure DinaBelova though does :) 16:28:31 my current task in my company is tracing request (API, RPC, DB, ...) --> find problems --> OSprofiler is the best candidate. 16:28:43 yay, /me here. almost here 16:28:47 tovin07_, neat :) 16:28:49 I evaluated OSProfiler with some services (Cinder, Glance, Neutron) and see a common workflow of OSProfiler: 16:29:05 (1) Services --> OSProfiler --> Messaging 16:29:06 (2) Ceilometer anotification pops out events from messaging (e.g. rabbitmq) and then process it --> panko --> MySQL 16:29:29 I see that, "Messaging" is now fixed in Glance. 16:29:46 (and other project too, such as Cinder) 16:29:48 #link https://github.com/openstack/glance/blob/48ee8ef4793ed40397613193f09872f474c11abe/glance/cmd/api.py#L80 16:30:11 harlowja_at_home, yes; still trying to figure out how trove should move forward with the oslo messaging issue. 16:30:20 not being able to replace the dispatcher is a drag 16:30:21 tovin07, by fixed you mean profiled? 16:30:31 the change that will result is significant 16:30:41 and no, moving to a rest api isn't the answer in the short term 16:30:46 tovin07_ in fact right now not only messaging/ceilo can be used :) but MongoDB as well. and (not released yet) several others like Elastic search 16:30:50 so I'm kind of between a rock and a hard place 16:30:57 I mean that, why we hardcoded it too “Messaging” 16:31:04 *to 16:31:30 that's not yet done in the all project :( 16:31:34 reviews are slow 16:31:47 all of the projects need to be updated like this https://review.openstack.org/#/c/316799/ 16:31:50 Recently, I see that you guys add some new drivers (redis, log_insight). However, I don't know how to switch to other driver and how that work. I've tried to change connection string in service config file (glance-api.conf, cinder.conf,...) and in `osprofiler show trace` --> no luck. 16:32:00 and everything will be grabbed from conf 16:32:40 I saw this one https://review.openstack.org/#/c/254703/ 16:32:41 amrith, right; a short term idea was using a custom serializer/deserializer, any attempts at that yet? 16:33:02 still working on it 16:33:16 And I know what you faced, DinaBelova: 16:33:43 amrith, ok, let's see how that goes, we can see if its really crap, if it is, then we can try a plan B of sorts 16:34:10 oh btw, anyone looking for a nice review to review, https://review.openstack.org/#/c/397264/ 16:34:26 julien trying to re-organize a bunch of hashring impls 16:34:35 Back to connection string, is there some magic here to use other driver (redis, mongodb,…)? 16:35:27 DinaBelova is the person that knows all the magic :) 16:35:38 sorry guys meeting in progress 16:35:42 cannot be really here 16:36:05 kk 16:36:18 tovin07, perhaps ping DinaBelova when she is free? 16:36:39 yes, I think so 16:36:42 :) 16:37:01 I see too much magic in that (OSProfiler) 16:37:29 I may be missing something obvious - but does anyone know where the source to this is: https://wiki.openstack.org/wiki/Oslo/Messaging 16:37:50 First hit when googling "openstack oslo messaging" (for me) 16:38:04 seems pretty stale... 16:38:25 kgiusti, ah, that, i think its just due for deletion/redirect to dev documents 16:38:50 kgiusti: it's a wiki, so you're looking at the source. 16:38:53 harlowja_at_home: cleaned up a bit - would be a nice base for a theory of ops 16:39:01 `Date of page creation 09:17, 29 April 2013` 16:39:09 so my guess it was the old dev docs 16:39:26 bknudson: thought so... :( 16:39:42 kk - I'll fireup the wayback machine.... 16:39:48 kgiusti, perhaps just keep the goals, concepts, references 16:39:55 and just redirect the rest to dev docs? 16:40:06 (and/or for more details go to dev docs) 16:40:07 +1 - 'zactly. 16:40:29 linking to or redirecting to some docs in git would be great. 16:40:47 that'd be fine with me, +1 to fixing these old docs that google finds 16:41:07 `In the degenerate case where none of the standard URL attributes have any meaning for a particular transport, we can do: 16:41:08 weirdo:///?param1=foo¶m2=bar` 16:41:09 lol 16:41:14 is that a new driver? 16:41:14 lol 16:41:27 +1 vote to rename amqp1 16:41:34 :) 16:42:59 alright folks, that's all i got, anything else anyone wants to bring up? 16:43:05 Something completely different: http://lists.openstack.org/pipermail/openstack-dev/2016-November/106906.html 16:43:14 ah, nice 16:43:31 I started this for o.m.: https://review.openstack.org/#/c/398578/ 16:43:34 thx for reminding about that kgiusti 16:43:55 `Stay tuned.` 16:45:04 oh, btw, since its thanksgiving week in the US, i expect people will not be around all that much :-P 16:45:12 people usually travel a bunch in the US this week 16:45:25 (if people are wondering where others are) 16:45:57 why were there centos-7 oslo.messaging jobs? 16:46:38 packaging for amqp1 driver was missing on ubuntu 16:46:39 was amqp1 driver only available on centos-7? 16:46:48 no longer the case 16:47:11 so don't need those anymore, great. 16:47:15 See https://review.openstack.org/#/c/399863/ 16:47:47 but begs the question: should *any* tests run on centos? Right now o.m. is strictly ubuntu. 16:48:06 (or will be once xenial transition is made - I assume) 16:48:20 hmmm, can't answer that, i like centos (it's what my work uses still) 16:48:38 (but we are also a few OS releases behind) 16:49:00 might want to ask infra if there's savings in not having centos jobs for oslo.messaging. 16:49:01 kgiusti: xenial transition is just to get off trusty,existing centos jobs are untouched 16:49:48 clarkb: kk - but currently all o.m. jobs that specify a node use ubuntu-trusty (and some xenial ATM) but no centos 16:49:58 Is this a blind spot in our testing? 16:50:34 (I'm speaking in particular about the driver tests - dsvm full + functional) 16:51:54 possibly? I cant answerthat for you 16:52:13 that was a question for the general audience :) 16:52:20 seems sorta like a blind spot? 16:52:46 * kgiusti would at least like to see the driver functional tests run against centos 16:52:59 IIRC they are pretty short in duration. 16:53:25 ya, i'm with u on that, seems useful (assuming its possible) 16:54:07 If there's no objection I can try adding just those as -nv jobs 16:54:12 wfm 16:54:19 for now 16:54:24 no objection from me :-P 16:54:59 kk I'll put together a quick patch for review... 16:55:09 cool beans 16:55:29 is there something special about centos that it needs different testing? 16:56:46 Not different testing - but some exposure is probably a good idea, the base systems (ubuntu/rhel) are fairly different, no? 16:58:23 guess we'll find out when we have that exposure? 16:58:59 (if not to useful, can just turn them off) 16:59:16 alright, running out of time :) 16:59:32 if needed #openstack-oslo for more :) 16:59:39 +1 16:59:55 and have a happy turkey day (for those who will be/are in the US) 17:00:06 #endmeeting