15:00:50 #startmeeting openstack-helm 15:00:51 Meeting started Tue May 8 15:00:50 2018 UTC and is due to finish in 60 minutes. The chair is mattmceuen. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:52 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:54 #topic rollcall 15:00:55 The meeting name has been set to 'openstack_helm' 15:00:58 GM everyone! 15:01:03 o/ 15:01:11 Here's our agenda: https://etherpad.openstack.org/p/openstack-helm-meeting-2018-05-08 15:01:20 please add in anything you'd like to discuss today 15:01:23 o/ 15:01:27 o/ 15:01:30 o/ 15:01:40 o/ 15:03:02 o/ 15:03:29 #topic Senlin, Mistral, Magnum 15:03:35 Thanks Pete :D 15:03:49 Any comments to share on those guys other than mission accomplished? 15:04:58 not really :) 15:05:10 lamt was a great help 15:05:19 for the senlin chart 15:05:19 thx lamt! 15:05:23 np 15:06:05 Next: 15:06:09 #topic release notes 15:06:32 At the PTG we had discussed the need for release notes as a gate for our 1.0 readiness (and moving forward with other releases) 15:06:54 I know that there has been some frustration with the way that Reno has been used / overused in some instances 15:07:12 However Reno's also the standard way of approaching release notes in openstack 15:07:39 I am open to any solution that saves developers effort and gets meaningful release notes (and no more!) into our users hands 15:07:52 As long as we use reno as a tool, not a weapon I'm good :) 15:07:59 As I don't have experience with Reno in a project, leaning on you guys to share opinions 15:08:47 o/ 15:08:49 Can you give examples of each portdirect? 15:09:01 hey srwilkers! Just in time to tell us how to Reno like Champions 15:09:15 you dont want me to tell you how to do that 15:09:18 yeah - I've seen reno gates that have really made the admin overhead of contibuting to a project very high 15:09:22 Reno like it's HOt 15:09:27 eg for a simple bugfix 15:09:44 exactly, reno like its hot -- throw ice water on it then throw it out the back 15:10:23 So to your point about using it like a tool -- 15:10:27 but then I've also seen projects where when used to describe and document major functionality changes or enchancement its been great :) 15:10:32 1) we need to define When we'd like to have release notes 15:10:42 2) we need to use something e.g. Reno to do only that 15:11:31 from the way the 1.0.0 spec is written - it looks like the versioning doc would be a good place to articulate this? 15:11:37 s/doc/spec 15:12:05 I think that's a good idea, it's very related 15:12:06 but yeah - if we get point 1 in place then i think we should be good 15:12:47 Cool. Any other Yeas / Neas on using Reno vs. an alternative, provided we have ground rules for whatever tooling we use? 15:13:13 nothing particular 15:13:45 There's a WIP for adding Reno: https://review.openstack.org/#/c/563481/ 15:13:57 Thanks to felipemonteiro_ for making it work :) 15:14:11 But I think we should hold off on merging it till we have ground rules defined 15:15:17 That's all I got on that topic 15:15:36 #topic Pike and Queens support 15:15:43 all yours portdirect 15:15:58 so - I've been doing some things on this side ;) 15:16:02 np, there's an infra follow-up if i recall correctly for the renos to be published 15:16:25 and have been running a full queens lab 15:16:39 and testing upgrade from newton to queens 15:16:47 and with running workloads.... 15:16:51 ...it works 15:16:54 :) 15:17:11 nice portdirect 15:17:13 yay! 15:17:14 +1 15:17:15 most of the changes i required are now in (or should hopefuly be in soon) 15:17:43 and in the next two days - I'll get checks in place for both of them 15:17:48 similar to how we do ocata 15:18:03 That is really fantastic. 15:18:05 though i need to test with kolla - as all my work has ben loci based. 15:18:37 https://review.openstack.org/#/c/566820/ >> this was only thing we found to do queens with kolla 15:18:42 pike works 15:18:52 yeah - loci did not have that issue :P 15:18:54 :D 15:19:13 thanks jawonchoo 15:19:28 https://github.com/portdirect/openstack-helm/commit/b3f82f3d3e6a912b88ccf7bc3a53e36377211487 15:19:32 np 15:20:04 To be clear, your test has been the full n->m->o->p->q, right? Is any stabilization between upgrade steps needed (beyond pods being ready) needed before moving on to the next upgrade? 15:20:14 nope 15:20:26 though my testing has been very simplistic 15:20:35 just launch a cirros vm at easch stage 15:20:43 and ssh into it, before moving on 15:20:46 neat 15:21:03 the hardest thing was cellsv2 support 15:21:21 Yeah that sucks portdirect - took me ages in kolla-k8s as well. 15:21:42 as to get `hands off` and be able to take advantage of them fully at later stages we need to do some footwork about their mangement 15:22:37 https://review.openstack.org/#/c/565457/13/nova/templates/bin/_db-sync.sh.tpl 15:24:42 Anything else on the version/upgrade front? 15:24:53 Is osh with loci images much quicker to deploy than with Kolla? 15:24:53 think thats it for now :) 15:25:42 rwellum: not really 15:25:45 rwellum: not much when we tested 15:25:51 you save a couple of minutes on the download 15:26:02 but frankly thats not super-relivent 15:26:17 what you really save on is simple builds, and improved security 15:26:50 Got it 15:27:41 #topic hands-on demo 15:27:48 take it away jayahn 15:28:17 i talked with pete 15:28:43 need communication and coordination needed as soon as possible. 15:29:28 i did "bold and underline" for the things decided among options. 15:29:48 * mattmceuen is reading it now 15:29:49 use newton, and upgrade to ocata (since this is the most tested and stable stuff) 15:30:10 basically prepare hands-on based on gate script 15:30:28 SKT will prepare VMs (100 VMs) 15:30:55 I had to do openstack-helm semniar to prepare the fund. :) 15:31:11 ah I get the underlining - cool 15:32:03 need to discuss scenario, espeically with showing lma. 15:32:33 does anyone else want to attend the meeting that jayahn is setting up on thurs? 15:32:49 i prettey sure we can deploy prometheus monitoring, but logging (with elasticsearch) requires quite a lot of resources. 15:32:49 would be great to get input from new people to the community as well. 15:33:07 +1. Please add me too 15:33:12 Me too 15:33:34 What time will it be? 15:33:37 pls put your name and email (contact info) on etherpad, so that I can invite you] 15:33:59 portdirect said he can be available either on thursday or friday. 15:34:18 so sometime in thursday. i will work on the exact time 15:34:24 yeah, using the full lma stack + osh on top for the workshops going to get resource intensive 15:34:26 thurday dring cst business hours is pretty much the only time i have this week :( 15:34:31 (since Friday there is Satureday here in Korea) 15:34:52 theres also a time constraint 15:35:16 so we'll need people to get the scripts to both be informative, and possible in the 1:30 we have 15:35:23 srwilkers: true. we can managee to deploy, but it will barely usalble. so my opinion is leaning toward just deploy prometheus and grafana. 15:35:57 I like that idea jayahn. We can always just tell people there's a full stack for logging as well. 15:36:00 jayahn: yep, that's the pretty dashboards anyway 15:36:06 And just deploy prom & grafana 15:36:08 yeap. 15:36:11 Exactly :-D 15:37:06 i will create etherpad entry to write things before the meeting, and invite all of you for thursday usa central time 15:37:57 https://etherpad.openstack.org/p/openstack-helm-vancouver-handson 15:39:34 okay. that's it for now. 15:39:41 Thanks jayahn 15:39:51 #topic Test Migration to StoryBoard 15:39:56 show us what you got rwellum 15:40:02 https://storyboard-dev.openstack.org/#!/project/195 15:40:08 It's a test migration. 15:40:14 nice! 15:40:20 Please check it out - and let me know what you think? 15:41:20 it this 'live' or we need to pull again if things change in launchpad? 15:41:21 certificaton warning. :) 15:41:48 portdirect: It's live but a test I believe. In other words, until you commit to cutting over it will get stale. 15:42:01 jayahn: really - not for me? 15:42:02 roger - looks ok 15:42:07 will poke for sure :D 15:42:25 Yeah sounds good - also look at Ironic's - they are quite far along. 15:42:35 will do 15:42:36 Awesome. Thanks for putting this together Rich. 15:42:47 I'll poke at this too 15:43:00 np 15:43:35 #topic PS needing review 15:43:45 There was a lot of happy stuff up above 15:43:54 But when it comes to PS needing review it is a sad sad day 15:44:05 All our PS from last week are still needing review 15:44:10 this is why i stay sad all the time -- i never get caught off guard 15:44:23 smart, srwilkers 15:44:40 I am as guilty obviously as anyone else :) 15:44:53 Let's focus on rounding these out in the next couple days please 15:44:56 https://review.openstack.org/#/c/559297/ 15:44:56 https://review.openstack.org/#/c/563224/ 15:44:56 https://review.openstack.org/#/c/543553/ 15:44:56 https://review.openstack.org/#/c/556325/ 15:44:56 https://review.openstack.org/#/c/560756/ > Need feedback on the last comment we put there. 15:45:41 This will insure maximum happiness in next week's meeting 15:45:56 * srwilkers schemes 15:45:58 Any other PS we need to call out for review? 15:46:46 #topic Roundtable 15:47:06 ok peeps - what else is on your minds? 15:47:39 As soon as Pete is happy with Queens support - I'd be interested in trying it out. 15:47:59 rwellum: next stop - master! 15:48:05 i probably shouldve thrown it in earlier, but the ldap stuff ive been working on for elasticsearch/kibana is almost ready for a proper review 15:48:21 need to tweak a few things, then can expect me to shout out for some eyeballs on it 15:48:22 +1 portdirect 15:49:37 Alright guys - good meeting :) 15:49:43 Thanks for all your effort 15:49:47 one quick side note: people tends to think osh is tight sync with loci, not kolla. we might need to do better communication on that. :) 15:50:01 but, we can chat later with this one. 15:50:40 we are free agents - the best tool for the job at hand ;) 15:50:52 yeap. that we need to let everyone know,. :) 15:50:57 That should be a talking point for us in Vancouver -- point out that we are image-agnostic to the extent possible 15:51:29 agreed 15:52:13 Meeting adjourned - see you all in the osh channel 15:52:16 #endmeeting