15:00:51 #startmeeting openstack-helm 15:00:52 Meeting started Tue Mar 27 15:00:51 2018 UTC and is due to finish in 60 minutes. The chair is mattmceuen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:53 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:56 The meeting name has been set to 'openstack_helm' 15:01:04 #topic rollcall 15:01:23 o/ 15:01:25 o/ 15:01:58 Agenda: https://etherpad.openstack.org/p/openstack-helm-meeting-2018-03-27 15:02:12 I'll give it a minute, because several of us in saint louis just got kicked out of our meeting room :) 15:02:35 Please add anything you'd like to discuss today to the agenda! 15:04:31 do we have jayahn in the house? 15:05:35 o/ 15:06:08 GM / GE everyone 15:06:43 #topic Multiline LMA functionality 15:07:20 This was a topic Jay wanted to discuss but I don't think he was able to join today. Unless anyone wants to cover for him? 15:08:07 We can table it till next time unless he is able to join in a bit. 15:08:19 That's a segueway into our next topic :) 15:08:28 #topic Team meeting time 15:09:13 At the PTG, it was brought up that our current meeting time is "morning in US" and "midnight in east asia", and opened the door to considering other timing. 15:10:31 I know not everyone is here (and timing is probably part of the reason), but how do we gathered feel about timing? Would a "afternoon US" time be good or bad? 15:11:49 i'd be fine with afternoon 15:11:56 `afternoon US` = evening europe 15:11:58 gmmaha - would that timing work for you? 15:12:04 afternoon/morning shouldn't be too much of an issue in the US 15:12:44 to portdirect's point - there's no time that will be good for everyone, but I think "evening for some folks" is probably better than "midnight for /anyone/" if possible 15:12:53 ++ 15:12:58 mattmceuen: yes afternoon would be fantastic. i don't have to run away in the middle of the meeting to drop the kid off at school.. 15:13:09 awesome :D 15:13:11 ++ 15:13:27 I'm down with that - long as we dont end up midnight in turkey etc 15:13:36 ^ 15:13:46 Good deal. Let me take an action item to set up a poll thingie with some potential times so the team can vote 15:15:08 #topic PS needing review 15:15:29 We have a few PS needing review, and they are probably some topic-lets in themselves 15:16:08 I took a stab at organizing the thoughts we'd had at the PTG around OSH 1.0 release requirements: https://review.openstack.org/#/c/556325/ 15:16:29 So that we can have broader discussion, nail them down, and then divvy them up as work items 15:16:52 If we can get some feedback / fill in gaps in that over the next week, it would be awesome 15:17:11 * mattmceuen pauses briefly to sip coffee before moving on to the next PS 15:17:45 Next is powerds0111's values ordering PS -- https://review.openstack.org/#/c/552485/ 15:18:07 There is some good discussion there so far, but I'd like for that to get broad input before we call it done 15:18:27 (spec I should say, not yet implementation) 15:19:08 o/ 15:19:13 bit late sorry 15:19:35 hey SamYaple no problemo! 15:20:33 Final PS that I think would be really good to get broad understanding/review of is the nginx sidecar PS: https://review.openstack.org/#/c/555386/ 15:20:56 do we have lamt present? 15:21:27 I think he may actually be tied up in a conflict today. 15:21:55 portdirect do you think this one is deserving of a quick spec to give context / promote discussion? 15:22:15 i can give it a review, we deploy nginx+uwsgi loci in production outside OSH 15:22:17 there should already be a spec i think 15:22:30 SamYaple: thats the next step on from this ps 15:22:44 as we need to remain compatible with Kolla 15:23:12 I'd like us to be able to have ` nginx+uwsgi` as an option (turned on by default) 15:23:24 but also support images without uwsgi 15:23:25 sure, i was more speaking to the configs needed to make it work 15:23:30 +++ 15:23:43 Spec: https://review.openstack.org/#/c/524013/ 15:23:55 :) 15:24:04 That still accurate for the new PS? 15:24:10 yup 15:24:20 awesome 15:24:42 Thanks in advance SamYaple - let's all give the PS some review 15:24:42 though we need a follow up for uwsgi 15:24:53 as this simply deals with stage one 15:25:07 getting nginx without uwsgi and nginx with uwsgi as deploy options might be... difficult 15:25:23 can nginx only do uwsgi? other options for non-uwsgi? 15:25:23 I've been giving it some thought 15:25:29 I think it shouldnt be too hard 15:25:36 ok cool 15:26:04 (mostly by going through home-salt... :D ) 15:26:28 sounds familiar 15:27:21 Any other PS that you all would like to get some wide feedback on? (or are otherwise languishing for review :-) ) 15:28:21 #topic Roundtable 15:28:34 Light agenda today - what else would we like to chat about 15:28:47 heat bootstrap? 15:28:49 versioning strategy? 15:28:57 templates in templates 15:29:09 ceph chart plans? 15:29:09 versioning strategy - someone is deep in speccing that out 15:29:25 SamYaple: I'm starting to pull together a versioning spec 15:29:43 Soon as I have a ps, you'll be the 2nd to know :) 15:29:47 got it 15:30:04 would be great for OSH and LOCI to follow the same one 15:30:15 (or variants thereof) 15:30:22 right, thats why im curious 15:30:28 i want to 1.0.0 15:30:28 I promise to have that up by next meeting 15:30:32 o/ 15:30:34 ++ 15:30:36 and plan to before EOW 15:31:01 thanks portdirect 15:31:15 gmmaha: we really need to split out the ceph chart into three (at least) 15:31:33 one for mons, another for osds, and one(or more) for fun extras 15:31:34 (do we have renmak_ -- he was interested in learning more about ^) 15:31:44 portdirect: right. sorry i have been out of action for a few weeks.. hence thought we shall sync 15:31:49 yes 15:31:53 I am here 15:32:04 which will allow is to follow the same upgrade workflow as ceph-deploy et al. 15:32:05 ohh hey renmak_ 15:32:13 I believe we should have user stories for those splitting up Ceph 15:32:35 we do *internally* ;) 15:32:49 :) 15:32:55 and we should have a bp really 15:33:20 is the ceph chart split 100% for ordering/upgrade reasons, or are there additional reasons too? 15:33:25 portdirect: i saw you had a wip for getting the mon chart out. guess that's a good place to start for now 15:33:34 seems like it might also be good to get ceph-helm people (like @rootfs?) to provide input on the design 15:33:58 mattmceuen: 50% upgrade 50% resiliency 15:34:18 d|k: we are doing this in line with sage's thoughts 15:34:42 the Arnold Palmer of production-grade software 15:34:45 fair enough, i just figured they might the heads up. 15:35:31 what kinds of fun extras would go in the third ceph chart? 15:35:42 rgw seems like one 15:35:45 mattmceuen: i am guessing rgw, rbd, etc 15:35:48 +1 15:35:56 ^^^ 15:36:07 in fact, rgw could be a component all by itself, i think. 15:36:22 hence - one (or more :) ) 15:36:25 yah 15:36:38 but lets go for: mons, osds, client things 15:36:42 as a 1st step 15:36:46 ++ 15:36:51 theres a LOT of work in that alone.. 15:36:56 there's probably also key management and mgr. 15:37:15 we also have to do this rapidly 15:37:27 as its def on the critical path for a 1.0 15:37:45 so anyone taking this on should be aware of that 15:37:58 portdirect: have you had a chance to write down your thoughts on how this will all playout?. 15:38:06 just to understand the design 15:38:50 guessing there will be a common ceph chart that will host keys, conf etc which we might inherit on all of the individual ones. 15:39:27 no! 15:39:33 lol 15:39:35 you cant upgrade that 15:39:38 :+ 15:39:45 we need them to be seperate 15:40:09 if we bundled them together, we'd have the same problem we have today 15:40:31 where we cannot upgrade each component individually 15:41:36 Pete, can you suggest/discuss design for this effort? It would be great to understand clear picture of what is needed 15:41:55 yeah, that will put us back to the where we are now. 15:42:55 I'll get a BP up 15:43:17 Awesome! 15:43:23 portdirect: thanks. sorry for loading more on your plate 15:43:32 np 15:46:52 Any other roundtable topics, or things you'd like to get on the agenda for next time? 15:48:47 cool beans -thanks everyone! have a good day 15:48:50 #endmeeting