14:39:36 #startmeeting CentOS Atomic SIG 14:39:36 Meeting started Tue Sep 30 14:39:36 2014 UTC. The chair is jzb. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:39:36 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:39:36 Meeting started Tue Sep 30 14:39:36 2014 UTC. The chair is jzb. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:39:36 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:40:00 #chair jbrooks scollier quaid walters kbsingh bexelbie 14:40:00 Current chairs: bexelbie jbrooks jzb kbsingh quaid scollier walters 14:40:00 Current chairs: bexelbie jbrooks jzb kbsingh quaid scollier walters 14:40:41 OK 14:40:48 so - going through last week's action items 14:40:51 the SIG proposal is up 14:40:58 We've appointed a chair 14:41:18 Have SIG folks who don't have wiki access made the request? 14:41:24 #info SIG proposal is posted on wiki page 14:41:27 I don't think I have an easy way to check that? 14:41:32 try to edit? 14:41:41 I think we use ACLs in some cases, so maybe just the SIG page 14:41:48 quaid: well, I meant to verify if all the members had asked 14:41:58 *I* have access. :_) 14:42:24 well, you can look to see if usernames are in the page source 14:42:35 Hm. OK 14:42:41 trick is as you recall one has to ask on centos-docs generally, and I didn't see any new ones that I recall this week 14:42:43 I'll shoot a note to SIG members this week 14:42:51 quaid: yeah, none are leaping to mind. 14:43:02 #action jzb ping SIG members on wiki access. 14:43:14 http://wiki.centos.org/Contribute#head-42b3d8e26400a106851a61aebe5c2cca54dd79e5 14:43:33 I still need to do a discuss on how Atomic should be listed in bugs.centos.org 14:43:34 #info SIG members should ask on centos-docs as per http://wiki.centos.org/Contribute#head-42b3d8e26400a106851a61aebe5c2cca54dd79e5 14:43:37 will do that today 14:43:51 #jzb open discussion on how Atomic should be listed in bugs.centos.org 14:43:54 er, 14:44:03 #action jzb open discussion on how Atomic should be listed in bugs.centos.org 14:44:15 walters: did you post on Anaconda work on atomic-devel? 14:45:07 i haven't yet sorry, haven't had a chance to test it in a clean CentOS environment 14:45:14 walters: ok 14:45:27 really i need a baremetal CentOS box, and I don't have one handy myself 14:45:43 walters: we can get you a free copy of CentOS... 14:45:45 though we could do the installer ISO in a virt environment 14:46:16 walters: are you short hardware? 14:46:46 walters, I've made my images so far in vms 14:47:05 i'd like to figure out a path for how Atomic works inside CentOS infrastructure 14:47:47 there are many components to that, from where the packages git lives, to the tree compose, and also the installer/cloud image generation side 14:48:38 walters: we can get you bare metal on *.centos.org I reckon 14:49:17 jbrooks, yeah...but i don't even want to contemplate trying imagefactory inside a VM, it's already painful enough =) 14:49:24 walters: I think there is a component in bugs.centos.org for 'build system', I think that's the best workflow we have to not forget stuff, then we'll get arrfab or KB to provision something for you 14:50:23 walters, ok -- I'll try it eventually -- is it documented somewhere, this imagefactory-based process? 14:51:16 jbrooks: we can also pester Ian McLeod, I think 14:52:21 OK 14:52:39 walters: you need a bare metal host to work with, yeah? 14:52:57 jzb, that would help yes 14:53:06 quaid: can you take that and get it provisioned? 14:53:24 quaid: or is that something I should request officially, or...? 14:53:27 jbrooks, it's now merged in https://github.com/projectatomic/rpm-ostree-toolbox - not really documented yet, still working on that 14:53:49 quaid: what's the process here for making the magic happen? 14:55:24 OK, we'll circle back 14:55:31 I'm going to give myself the action to pester quaid 14:55:51 #action jzb to follow up with quaid on getting hardware for bare metal work under *centos.org 14:57:03 OK 14:57:09 (quiet today) 14:57:17 bexelbie: IIRC you have some questions on docs 14:57:23 bexelbie: you want to take that? 14:57:32 sure 14:57:53 sorry, was in another window 14:57:53 I was looking at putting the Atomic doc up and was wondering if we could keep it in a git repo instead of the wiki 14:58:12 I think the process is as above, file a work ticket in bugs.centos.org in the build system component and ask for a bare metal host for Atomic work 14:58:14 I think for long term maintainability it is a better plan and it will create what I think is a better workflow for future docs 14:58:46 bexelbie: which doc? best workflow somewhat depends on who will collaborate on it 14:59:57 the first doc is the Atomic QSG/background info 15:00:18 git would seem to open it up to good revision control for future contributions from my side and allow full community access 15:00:35 +1 to git 15:00:39 I realize wiki's do revision control ... 15:00:56 +1 to git as well 15:00:56 however, wiki docs tend toward chaos in my experience and it isn't clear we have a lot of wiki garderners around 15:01:09 I am thinking longer term here, so in the short term this won't crop up at 15:01:26 It will also make future contributions from my side mucho easier :) 15:01:27 if the goal is for the SIG to work on it, then whatever the SIG prefers; keep in mind that is a barrier to others contributing, that's all 15:01:55 I believe that git is a barrier that can be overcome - that said I also know of some tools in progress that may help us with that in the long run 15:02:01 I look to the SIG for guidance :) 15:02:33 jzb: yeah, open a ticket in bugs.centos.org, should be a build system component, and fill it with the details -- including how long the provision is for 15:02:34 Personally, I feel much more comfortable sending a PR vs rewriting stuff 15:02:40 quaid: OK, will do 15:02:50 walters: how long do you expect to need it? 15:02:59 walters: 1-2 months seem sufficient, or is this longer term? 15:04:03 I like PRs as well as it allows a doc to have some editorial control and can illuminate things that need second reviewers more easily 15:04:11 jzb, as long as it takes for the process to be more integrated with the centos rel-eng; e.g. if we can use imagefactory-as-a-service instead of imagefactory locally, that helps avoid the need for a baremetal box 15:04:13 wikis are generally alwasy a "commit = publish" system 15:05:37 bexelbie: I think we are generally disposed towards git, can you send a note to centos-devel? 15:05:49 bexelbie: or maybe centos-docs 15:06:02 jzb, sure - do you want a proposal for this and ask for location? 15:06:25 bexelbie: pretty much, put out the workflow as you see it working and see if there's any dissent / better ideas, etc. 15:06:44 jzb, sounds like a winner - I'll put it on my todo for this week 15:06:49 bexelbie: thanks! 15:06:51 bexelbie: I'd recommend (from a doc perspective) to use the best tool for the team (FLOSS, of course) to get things done over the shorter term, and to simply put it on a list of things to review, "Are we now where it makes more sense to move to an easier-to-edit-by-the-masses solution?" 15:06:53 walters: ack 15:07:20 OK, that's the items that I have for today, other items? 15:07:31 quaid, ack - I have a dream about how we define "Easier to edit by the masses" :) we can talk about that later 15:07:43 #action bexelbie put out docs workflow to mailing list to see if there's any dissent/better ideads. 15:07:46 er, ideas. 15:08:05 #topic new business 15:08:08 jzb, I'll start with centos-docs 15:08:12 dumb question, but is walters outside expert or a SIG member? 15:08:46 quaid, i'm not a sig member at the moment, but i am obviously interested =) 15:08:53 quaid: I think he's a pending SIG member. Needs to "officially" request 15:09:13 walters: all you really need to do is send a request to centos-devel so we can approve SIG request 15:09:15 in short, resources would be for SIG members, so maybe as part of this just complete that loop so the governance/access is all aligned 15:09:21 +1 15:09:37 walters: slightly more formal than Fedora, I suppose :-) 15:10:53 any other topics or items for today? 15:11:06 looking 15:15:54 I think walters is digging deep :-) 15:16:05 :) 15:16:53 sorry back; sent the request 15:17:18 walters: thanks! 15:18:03 to what degree are we thinking of the Fedora Atomic efforts as upstream? 15:18:36 walters: that's a tricky question. ATM, I think concurrent 15:19:50 right. There's a lot of overlap of course, main thing is to ensure we don't diverge unnecessarily on things like package set 15:20:02 walters: so, I'm working on that :-) 15:20:15 walters: by end of week I hope to put out a proposal with the package set 15:20:18 walters: I think you've seen that. 15:20:21 early draft 15:20:26 yep, cool 15:21:00 #action jzb get Atomic definition out by end of week to Project Atomic (SIG will want to review) 15:21:06 walters: thanks 15:21:10 any other items this week? 15:22:54 I'm going to take that as a no :-) 15:23:07 thanks everybody, this isn't goodbye, it's just "until next week" 15:23:12 see you on the mailing lists! 15:23:15 thanks all! 15:23:16 #endmeeting