14:00:13 #startmeeting airship 14:00:14 Meeting started Tue Oct 8 14:00:13 2019 UTC and is due to finish in 60 minutes. The chair is nishantkr. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:15 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:17 The meeting name has been set to 'airship' 14:00:24 #topic Rollcall 14:00:30 o/ 14:00:37 o/ 14:00:37 o/ GM+GE 14:01:06 Hello everyone ! so far we have no agenda items, so please add anything you would like to discuss here - https://etherpad.openstack.org/p/airship-meeting-2019-10-08 14:02:01 May be a light crowd to go along with our light agenda, a number of folks are travelling or ill today 14:02:20 yup looks like 14:03:03 you lucked out on your day to host nishantkr ;) 14:03:11 #smart 14:03:11 haha 14:04:24 since we have no agenda items to discuss specifically, If anyone would like to discuss on what they are working on and would like to share it here then please do, I can go first 14:04:38 sounds good to me 14:05:31 I am working on efforts for Airship to move towards kubernetes 1.16.0 14:05:43 Here are the PS out there for the same - https://review.opendev.org/#/q/topic:k8s-1.16+(status:open+OR+status:merged) 14:05:44 ++ 14:06:07 are those ready for some review, nishantkr? 14:06:22 yup some of them are ready for review 14:06:34 I am doing another round of testing 14:06:47 on the same note, k8s-entrypoint is moving to 1.16. I had to push another PS to force the gates to kick off, so its +2's are gone :( 14:06:56 PS here https://review.opendev.org/#/c/686829/2 14:07:05 do we need to make any changes to the version of Helm we are using for K8s 1.16? 14:07:43 nope i did not make any changes to helm version and i was able to get a succesfull run of virtual seaworthy 14:07:44 dwalt: possibly, there was some weirdness with tiller. Let me find the PR 14:07:52 sounds good howell 14:08:06 o/ 14:08:48 o/ roman_g 14:08:52 I remember someone told that docker should be upgraded to use newer k8s 14:08:59 Scott may be 14:09:04 here we go, not sure if this actually affects us though: https://github.com/helm/helm/pull/6462 14:09:21 I mostly followed the release notes and deprecated notes for each version starting 1.13 14:09:24 https://kubernetes.io/docs/setup/release/notes/#deprecations-and-removals 14:10:08 That issue won't affect airship 14:10:09 howell: ah, I see. So we are okay then since we have our own Tiller chart? 14:10:12 As we don't use helm init 14:10:27 Majority of the changes are related to charts using apiversions which are deprecated and not supported in k8s 1.16 14:10:42 Along with some kubelet params which are deprecated as well 14:10:46 The only gap I can think of is Airskiff - but we use the OpenStack-Helm minikube scripts. They're also making the transition to K8s 1.16 already. 14:10:50 But we should likely upgrade tiller as any new APIs in 1.15 or 1.16 likely won't be supported in Helm 1.14.x 14:12:15 though it doesn't look like there is a helm 2.x release supporting Kubernetes 1.16 14:12:30 hmm that's unfortunate 14:12:35 I think followed by this change we are also planning to move towards dynamic kubelet config, not sure if anyone has already started working on it 14:12:44 Jagan Mohan Kavva proposed airship/porthole master: Chart/Dockerfile for etcdctl Utility Container https://review.opendev.org/674680 14:13:31 I'm not sure if that is a hard drive to helm 3.x or indicates kubernetes hasn't added any net new API resources since 1.14 14:13:44 `helm init` was broken for kubernetes 1.16, but airship doesn't use that. I'm not aware of any other issues with e.g. helm 2.14.3 against kubernetes 1.16 that have been raised 14:14:42 but yeah would need to update tiller for kubernetes 1.16 client lib once that's released 14:15:07 Yup seaneagan 14:15:23 Any other questions on this topic? 14:15:24 What is the stance from Airship on components expected to be retired for Airship 2.0? 14:15:38 Sorry, ^ is another topic unrelated 14:16:11 no problem sthussey, we are going in an adhoc manner today 14:17:09 plan is to have an upgrade path and/or migration period where both are supported 14:17:37 don't think we have discussed about that yet for individual components but there was some discussion around treasuremap and i think we are going ahead with creating separate directories for Airship 2 documents 14:18:02 Someone has inquired about the MAAS chart as an isolated use-case 14:18:18 how long do we plan to support airship 1.0 components after 2.0 is released? 14:18:22 But I can't in good conscience recommend they use it if it will go into cold storage in less than 12 months 14:18:34 yeah, for treasuremap, there is a plan to create a 2.0 specific directory; and a separate branch for that directory until it gets off the ground 14:19:24 Something like a per-component roadmap may be worthwhile 14:19:44 Around how long things like security fixes, gate maintenance, etc... will go on 14:19:57 that's a good point sthussey. If folks are still getting "non-airship" use out of airship widgets, I have no problems with leaving them available (albeit potentially with less support). I think Plan A would be to encourage folks to migrate to Airship 2.0 though, in part or in whole, when the time comes 14:20:14 or, working with others to find new homes for projects -- like we did with k8s-entrypoint 14:21:00 jamesgu__: I don't think that's been formally determined beyond "only as long as we need to". I think that's still in "cross that bridge when we get to it" territory 14:21:36 the idea is to not support two Airships, though, and provide a smooth migration path, so that should minimize the time spent supporting two versions substantially 14:22:12 That only makes since in terms of a monolith. There is no migration path if your use case is 'deploy Canonical MAAS inside Kubernetes' 14:22:20 agree 14:22:20 Merged airship/kubernetes-entrypoint master: Bump Kubernetes version to 1.16. https://review.opendev.org/686829 14:22:27 A cluster-api operator using Ironic isn't going to help 14:22:59 I'll just send this fellow to the discussion list, he can work with the directional committees 14:23:13 yup, thanks for bringing that topic sthussey, it's something we need to discuss and determine formally 14:23:26 although I would encourage the maas team to integrate with Metal3 / Cluster API as well to remain a compelling in-k8s provisioner 14:24:39 ok i do see one agenda item from pramchan 14:24:48 #topic Any preferred IDE for go airship team uses 14:25:21 speaking for the entire team: vim, full stop 14:25:33 * mattmceuen ducks 14:25:41 haha 14:26:10 vim does have a go plugin :-) 14:26:34 I won't throw anything at you for that one mattmceuen :) 14:26:58 howell goes the most these days, what do you use howell? 14:27:49 in any case though, we don't want to tie airship to any ide in any way shape or form 14:28:27 as far as vim plugins, I use vim-go and ALE. They work together pretty nicely 14:28:33 Yes I am reviewing your comments 14:28:43 awesome 14:28:56 maybe I was speaking for the team after all 14:29:16 Jagan Mohan Kavva proposed airship/porthole master: Chart/Dockerfile for etcdctl Utility Container https://review.opendev.org/674680 14:29:40 anything else you would like to discuss on that pramchan? 14:30:43 Sure any one has any template for airship project for Komodo editor? 14:31:06 For Go specific , I was trying to use 11.1 version 14:32:41 I've never even heard of that editor, sorry pramchan 14:33:04 goland is pretty popular in general go dev 14:33:06 One more my teams wanted to use Intelli-J for IDE , I just wanted to see what is the best opensource editor besides vimeo or vi-go ALE mentioned above. 14:33:09 not free though 14:33:09 it seems like the go community has settled on vim, vscode, and goland as he primarty editors 14:34:37 if you want to get really close to go, you can use 'acme' which is the preferred editor of Rob Pike, creator of Go 14:36:00 Thanks then I will settle for one of above what you use, unless Komodo and IntelliJ teams 14:36:16 warning about acme: it doesn't use a monospace font 14:36:23 at least, not by default 14:36:40 one thing to add to our ad-hoc agenda today: review requests. It's a mix of ones from last week that are still outstanding, plus reiterating the k8s 1.16 PS that were mentioned up above 14:36:40 Decide to justify some compelling reasons to use their tools 14:36:44 Requests for review: 14:36:44 Leftovers from last week! 14:36:44 https://review.opendev.org/#/c/676700/ - Spec: Introduce isogen subcommand for airshipctl 14:36:45 https://review.opendev.org/#/c/675851/ - Airshipctl: Add logic to isogen subcommand 14:36:46 https://review.opendev.org/#/c/679563/ - Airshipctl: Generate cloud init settings 14:36:47 https://review.opendev.org/#/c/685049/ airship/docs Add Ansible code formatting documentation 14:36:48 https://review.opendev.org/#/c/683209/ pegleg add support for global encryption 14:36:49 New review requests: 14:36:50 https://review.opendev.org/#/q/topic:k8s-1.16+(status:open+OR+status:merged) - k8s 1.16 14:36:51 https://review.opendev.org/#/c/686829/ k8s-entrypoint to 1.16 14:36:54 great, thanks pramchan hopefully you got some answers 14:37:21 Did and thanks will use the guidance. 14:37:27 yup please take out some time and review those PS 14:37:36 #topic Roundtable 14:37:51 Anything else someone would like to discuss/share? 14:39:10 I'm "YAML programmer", I have no preference of text editor and can work with pretty much all of those popular ones 14:39:25 /sarcasm 14:39:27 lol 14:39:40 ok good meeting guys from our almost "zero" agenda. 14:39:44 Thanks for joining 14:39:56 #endtopic 14:39:57 Thanks everyone :) 14:40:02 thanks nishantkr and team! 14:40:02 #endmeeting