16:02:29 <jzb> #startmeeting CentOS Atomic SIG 16:02:29 <centbot> Meeting started Thu Jan 22 16:02:29 2015 UTC. The chair is jzb. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:02:29 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic. 16:02:47 <jzb> #chair jbrooks kbsingh Evolution imcleod 16:02:47 <centbot> Current chairs: Evolution imcleod jbrooks jzb kbsingh 16:02:59 <jzb> I don't see nzwulflin here 16:03:06 <jzb> who else have I missed? 16:04:00 <kbsingh> colin ? 16:04:04 <jzb> ahhhh 16:04:09 <jzb> #chair walters 16:04:09 <centbot> Current chairs: Evolution imcleod jbrooks jzb kbsingh walters 16:04:29 <jzb> OK 16:05:00 <jzb> imcleod: anything you'd like to report this week? 16:05:21 <imcleod> jzb: Hiya 16:05:51 <imcleod> jzb: I have made a few runs at building Vagrant and qcow2 images from the nightly content, thus far without success. Am still working on root cause. 16:06:03 <imcleod> jzb: I'm generally upbeat about it, just need to find the time to find the problem. 16:06:25 <jzb> imcleod: anything we can do to help there? 16:06:56 <imcleod> jzb: I suspect not. 16:07:13 <jbrooks> imcleod, Is this through koji? 16:07:49 <imcleod> jbrooks: It is indeed. One of the problems, unfortunately, is that when the build fails in koji, the screenshot we get is not helpful, because of some bug with the text output/screensaver. 16:08:00 <imcleod> jbrooks: SO I have been doing root cause on another machine using Factory directly. 16:08:11 <imcleod> jbrooks: But the needed bits are in the CBS koji and should work once we have this sorted. 16:08:21 <imcleod> jbrooks: I have used them to build non-Atomic images. 16:08:31 <imcleod> jbrooks: And if you want to fiddle with them, that would be fantastic. 16:08:41 <jbrooks> imcleod, if your steps are documented, yeah, I can fiddle w/ it 16:09:05 <imcleod> jbrooks: What I should probably do is verify that the issue can be duplicated using rpm-ostree-toolbox. rangerpb would love it if I did that. 16:09:46 <kbsingh> fwiw, rpm-ostree-toolbox isnt a happy place st the moment either 16:10:08 * kbsingh will update when his turn comes 16:11:20 <jzb> imcleod: is that it on your update? 16:11:30 * jzb should remember to do "topic" 16:11:39 <imcleod> jzb: That is all for me, I think. 16:11:48 <jzb> imcleod: groovy, thanks 16:11:55 <jzb> #topic update from kbsingh 16:12:02 <jzb> kbsingh: it's your turn :-) 16:12:44 <kbsingh> so, we ran the nightly builds for a few days ( a week actually ) 16:12:55 <kbsingh> for ostree repo, and the lorax tree's 16:13:39 <kbsingh> however, 2015.1-1.el7 seems to have an issue where it does not like being run from Cron anymore 16:14:02 <kbsingh> it fails to do chroot into the new filesystem its setting up, and i need to go look at that sometime 16:14:07 <jzb> kbsingh: that's rpm-ostree? 16:14:13 <kbsingh> ( this is background to my note to centos-devel list btw ) 16:14:30 <kbsingh> jzb: yeah, well - ostree and ostree-toolbox, same ver, updated at the same time 16:14:37 <jzb> kbsingh: gotcha 16:15:01 <kbsingh> I spoke with walters about this at the time, and he had a look at logs etc - we fixed a bunch of other things, but were unable to resolve this specific issue 16:15:15 <kbsingh> works fine from cli, not so much from cron, so its almost certainly going to be an ENV issue 16:15:38 <jzb> kbsingh: so you're having to do it manually, then 16:15:43 <walters> kbsingh, maybe selinux policy? 16:15:49 <kbsingh> need to find the time to go investigate a bit; in the mean time I've kicked builds manually a few times, so none running centos atomic host is exposed 16:15:55 <walters> any avc denials? 16:16:15 <jzb> kbsingh: thanks 16:16:21 <kbsingh> selinux is actually disabled on the host, i think the ostree-2014 builds needed it turned off 16:16:56 <kbsingh> that was 50% of my update 16:17:17 <kbsingh> the other 50% is that I've been pushing the atomic host builds at EC2, and now have AMI's in every region based on the December build 16:17:49 <kbsingh> whatever warts and bugs it comes with, its there - and fr the Jan 29th build, can we consider telling people about these ? 16:17:55 <kbsingh> we dont need to decide right now 16:18:21 <jzb> kbsingh: yes, we should tell people about those 16:18:22 <kbsingh> but the push-atomc-images-into-ec2 task is plumbed into the gen-centos-cloud-images-vendor process 16:18:22 <jzb> them 16:19:06 <kbsingh> <end of update> 16:19:14 <jzb> kbsingh: awesome, thanks! 16:19:29 <jzb> walters: did you have anything to add this week? 16:20:20 <walters> just that whatever we can do to move forward with anaconda would be great. I should be able to help with patch merging 16:20:27 <walters> one pain point is the lack of dist-git 16:20:42 <walters> because we have to continually merge the centos anaconda branding patches with the atomic patches 16:21:30 <kbsingh> how can we make that easier 16:21:55 <imcleod> walters: I should be doing that. Apologies. Do you expect the most recent downstream update to work if pulled in? 16:22:11 <imcleod> walters: If so, I'll grab it right now, reapply those few centos patches and kick off a CBS build. 16:23:18 <walters> kbsingh, dist-git would be a collaboration point, although we could just import the srpm into projectatomic git 16:23:33 <walters> i'm fine with the latter, just need to make a decision on where we keep the source 16:23:52 <kbsingh> where is projectatomic git ? 16:24:04 <walters> https://github.com/projectatomi 16:24:05 <walters> c 16:24:07 <kbsingh> we can host stuff on git.centos.org as well, with the free koji integration that would/could come from it 16:26:04 <jzb> kbsingh: can we just pull to git.centos.org from github? 16:26:07 <jzb> or push to? 16:26:38 <jzb> kbsingh: the reason I ask is we might get more contributions from folks on GitHub 16:26:41 <jzb> (might 16:26:59 <jzb> either way - let's make a decision today and go forward 16:27:15 <jzb> this seems to have come up several times and I think even a "bad" decision is better than no decision. 16:28:48 <kbsingh> so, i just checked.. 16:29:06 <kbsingh> what we can do is setup a mirror inbound, and use that to build into koji from - but someone needs to take ownership for the mirror 16:29:26 <kbsingh> eg: imcleod might take it up to sync git.centos.org with content from github.com, and then request koji builds for the components 16:29:32 <jzb> imcleod: ^^ does that sound like you? 16:29:49 <imcleod> jzb: I'll give it a try. 16:29:53 <kbsingh> it can be a number of people, but it just needs to be a set of people who's details are setup at git 16:30:09 <kbsingh> ( as opposed to : running a mirror where anything on github automagically makes it to git.c.o ) 16:30:41 <jzb> kbsingh: totally reasonable. 16:30:47 <imcleod> I don't believe I'm set up on g.c.o yet. 16:30:49 <imcleod> So I should do that. 16:31:20 <kbsingh> we can do some trial runs at Fosdem ( am hoping to do this with hhorak and the scl folks too ) 16:31:23 <jzb> kbsingh: how does one get set up on gco? 16:31:34 <jzb> kbsingh: I'm seeing a login but not an obvious signup 16:31:49 <kbsingh> need to ask via bugs.centos.org includ sig name ( same as koji ) 16:32:35 <kbsingh> unified central auth will make this redundant, a koji account gets you git automatically 16:33:45 <jzb> kbsingh: got it. 16:33:56 <jzb> kbsingh: one bug req per person or can we req several at once? 16:35:28 <jzb> imcleod: Imma give you the action to be Mr. Mirror, OK? 16:35:43 <imcleod> jzb: Sure. Imma let you do that. 16:36:17 <jzb> awesome 16:36:49 <jzb> #action imcleod to request git.centos.org credentials, work on acting as "mirror" for github/atomic -> git.centos.org 16:37:05 <jzb> walters: do you want to set up a repo in github for this work? 16:37:10 <jzb> under the Atomic project? 16:38:43 <kbsingh> jzb: i think one per person is best, just easier to track 16:39:03 <jzb> kbsingh: ack, OK 16:39:17 <kbsingh> i think there might be quite a few repos, one per package ? 16:39:17 <jzb> might have lost walters to other stuff 16:39:46 <jzb> #action jzb work with walters and imcleod to set up github repo for collaboration 16:40:19 <jzb> #topic Activity Day at SCALE, Project Atomic at Devconf 16:40:39 <jzb> quick update - we're doing a Fedora Activity Day at SCALE around cloud + Atomic 16:41:14 <jzb> I know that's Fedora-specific-ish, but I suspect there's some overlap w/CentOS so I wanted to mention it here and welcome anybody (imcleod, Evolution) who might want to attend. 16:41:18 <jzb> it's on Sunday 16:41:32 <jzb> we're also doing a workshop on Saturday at DevConf.cz from lunch till end of day 16:41:36 <jzb> for Project Atomic 16:41:45 <jzb> I hope walters will be able to make that. 16:41:48 <Evolution> jzb: sure. can be there. I don't fly out until monday (scale) 16:41:55 <jzb> Evolution: excellent 16:41:55 <Evolution> devconf I can't do. 16:42:07 <jzb> Evolution: Yeah, I knew you were heading home from that 16:42:23 <jzb> that was my piece for this week 16:42:28 <jzb> #topic open floor 16:42:45 <jzb> any other topics this week? I think most folks are a bit slammed with other things, like getting ready for FOSDEM. 16:42:50 <jzb> nzwulfin: hey! 16:42:55 <kbsingh> I'm at devconf on Sat, will stop by 16:42:59 <jzb> nzwulfin: I see you made it 16:43:03 <jzb> nzwulfin: anything for this weeek 16:43:04 <jzb> ? 16:43:07 <jzb> kbsingh: awesome 16:44:48 <jzb> anybody? 16:44:49 <jzb> going once ... 16:45:23 <kbsingh> one thing.. 16:45:34 <kbsingh> I've got 2 physical machines in ci.centos.org earmarked for atomic tests 16:45:35 <jzb> kbsingh: OK 16:45:40 <kbsingh> but i have no tests 16:45:55 <kbsingh> if anyone has tests floating around, do tell :) 16:46:12 <kbsingh> was hoping to work on this a bit early next week, but other firefighting going on 16:46:18 <jzb> kbsingh: perhaps we can work on that during the workshop in Brno 16:46:27 <kbsingh> I'm happy to demo etc at Dojo/Fosdem /devconf 16:46:35 <kbsingh> sounds good 16:46:51 <jzb> kbsingh: thanks! 16:48:47 <imcleod> Heh 16:48:51 <imcleod> sww 16:48:58 <jzb> OK 16:49:01 <jzb> calling it - thanks everybody 16:49:05 <jzb> #endmeeting