17:00:53 #startmeeting atomic-wg 17:00:53 Meeting started Wed May 10 17:00:53 2017 UTC. The chair is dustymabe. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:53 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:53 The meeting name has been set to 'atomic-wg' 17:00:57 #topic rollcall 17:01:04 .hellomynameis kushal 17:01:05 kushal: kushal 'Kushal Das' 17:01:05 .hellomynameis dustymabe 17:01:07 dustymabe: dustymabe 'Dusty Mabe' 17:01:09 .hello yzhang 17:01:10 .hello jberkus 17:01:11 yzhang: yzhang 'Yu Qi Zhang' 17:01:14 jberkus: jberkus 'Josh Berkus' 17:01:19 .hello jasonbrooks 17:01:20 jbrooks: jasonbrooks 'Jason Brooks' 17:01:45 .hello sayanchowdhury 17:01:46 sayan: sayanchowdhury 'Sayan Chowdhury' 17:01:50 .hello miabbott 17:01:53 miabbott: miabbott 'Micah Abbott' 17:02:07 .hello roshi 17:02:08 roshi: roshi 'Mike Ruckman' 17:02:58 #chair jberkus jbrooks sayan miabbott roshi yzhang kushal 17:02:58 Current chairs: dustymabe jberkus jbrooks kushal miabbott roshi sayan yzhang 17:03:06 #topic previous meeting action items 17:03:18 ok did we have a meeting last week or did we cancel it ? 17:03:30 sorry for not being there, was busy with conference stuff 17:03:32 There wasn't one 17:03:43 As far as I can remember 17:04:04 ok 17:04:10 will find AI from previous meeting 17:04:11 no meeting last week 17:04:21 * jbrooks to lead writing up the new support policy and plan for rolling 17:04:24 jbrooks: ^^ 17:04:37 https://pagure.io/atomic-wg/issue/228#comment-439326 17:05:10 That can become a blog post or something 17:05:29 pending feedback and whatnot 17:05:42 jbrooks: reading now 17:05:48 bear with me 17:06:27 Hola 17:06:57 jdoss: hiya 17:07:01 #chair jdoss 17:07:01 Current chairs: dustymabe jberkus jbrooks jdoss kushal miabbott roshi sayan yzhang 17:07:14 howdy all 17:07:16 jbrooks: that looks mostly good to me - maybe one nit - i'll comment in the issue 17:07:18 .hello jdoss 17:07:19 jdoss: jdoss 'None' 17:07:22 cool 17:07:25 jzb: hey 17:07:28 #chair jzb 17:07:28 Current chairs: dustymabe jberkus jbrooks jdoss jzb kushal miabbott roshi sayan yzhang 17:07:29 .hellomynameis jzb 17:07:30 jzb: jzb 'Joe Brockmeier' 17:08:04 any other comments for jbrooks on that action item? 17:08:14 dustymabe: I have something to discuss regarding fedora cloud 25 and the lack of fedora cloud 26 alpha Vagrant/EC2 test images. 17:08:33 jdoss: ok, we have open floor for that 17:08:37 towards the end of the meeting 17:08:42 sound good? 17:08:49 Kk sounds great 17:09:09 #topic fedimg: don't use 'builder' instance for uploading AMIs 17:09:14 #link https://pagure.io/atomic-wg/issue/269 17:09:28 jbrooks: I think we'll want to add something about how support for N-1 is "best effort", as in not tested at all 17:09:33 even before we go to rolling updates 17:09:46 sayan gave some of us an overview of fedimg yesterday 17:09:55 it's in the text, but I think it needs to be its own bullet point 17:10:30 sayan: also gave us an update on the status of this ticket 17:10:36 right now I am deploying the changes into dev instances to test how things are working 17:10:51 jberkus, bullet 4 is solely about that, but I could repeat it at the end 17:10:53 sayan: great! 17:10:55 http://ec2-184-72-150-211.compute-1.amazonaws.com/ami/jobs/3/output 17:11:23 sayan: so that is a failure? 17:11:31 jbrooks: more I think we need to more directly spell it out. I'll take a stab at it 17:11:32 dustymabe: yes, that's a broken AMI 17:11:41 sayan: ok 17:11:59 sayan: so one other thing, there is some other stuff pending on the libcloud update. can you give karma to that update? 17:12:36 dustymabe: regarding that issue, the maintainer did not build for el7 17:13:00 I dropped him an email few days back, and I was told that he would be updating that for el7 too 17:13:20 sayan: and you only use it on el7? 17:13:27 dustymabe: yes 17:13:46 ok 17:13:51 thanks sayan 17:14:01 next topic 17:14:16 I am planning to a scratch build and send over to the maintainer so that it's easier for him 17:14:24 #topic clarify policy on atomic host support for older Fedora "number" releases 17:14:29 #link https://pagure.io/atomic-wg/issue/228 17:14:39 this is what jbrooks talked about earlier 17:14:48 see https://pagure.io/atomic-wg/issue/228#comment-439326 17:15:06 please look at that issue and comment on his suggestion 17:15:13 anything else? 17:15:40 . 17:15:50 moving on... 17:15:58 #topic design, deploy and document Fedora OpenShift Playground (FOSP) 17:16:07 #link https://pagure.io/atomic-wg/issue/153 17:16:17 same state? 17:16:20 dustymabe: blades are getting racked. no clear timeline on when FOSP will be available to us 17:16:26 dustymabe: question for the group 17:16:47 i'm with the fedora infra crowd this week at a FAD 17:16:54 anything you want me to ask any of them specifically ? 17:16:54 dustymabe: given the delays in making Openshift v3 available, I'm reconsidering whether we want to have a "playground", or whether we want running FAO hosting 17:17:19 dustymabe: is fedora infra involved? or is this all OSAS infra? 17:17:29 no idea 17:17:39 I think openshift v3 is available now 17:18:12 what does v3 have anything to do with it? 17:19:19 w/o v3, there's no public place to play w/ openshift 17:19:44 jbrooks: are you referring to openshift online? 17:19:51 i'd just request that the deployment is on bare metal, or at least some of it is - this would help us move some of our VM-based tests into it 17:19:52 openshift v3 has been out for a long time 17:20:09 No 17:20:20 Well, what do you mean by a long time? 17:20:24 We're talking online 17:20:27 dustymabe: I'm referring to v3 online 17:20:35 ok yeah 17:20:40 please don't refer to it as just v3 17:20:45 that has no meaning 17:20:48 dustymabe: I need somewhere to host the new pa.io site, among other things, and I'd like us to EYODF, which means FAO 17:20:56 heh, I knew what he meant 17:21:10 well FAO would run openshift v32 17:21:12 v3* 17:21:21 right, but we need HW for hosting 17:21:33 right - ok we're all on the same page 17:21:35 jberkus, yeah, I'm on v3 online now -- just signed up 17:21:44 I'm suggesting that we dump the idea of a "playground" and instead have a hosted instance of FAO we can use to host things 17:21:51 jbrooks: you are? I'm still waiting 17:22:08 where is a hosted instance of FAO going to be hosted? 17:22:40 jberkus, I am waiting for openshift.io, which I tried to sign up w/ w/ my rh acct, and that's blocking my access to regular v3 online. I just tried signing up w/ my personal acct, and that worked 17:23:17 dustymabe: the plans for FOSP was to host it in the OSAS cage 17:23:25 Annoyingly, if you're waiting for openshift.io, you're blocked for online, since they don't want you double-dipping 17:24:08 ok i thought you were sugguesting we dump the OSAS stuff 17:24:37 either way, we're talking in circles. status: still waiting on hardware, but we may be able to find hosting for pa.io in openshift online before we get FOSP 17:24:51 i'll move on to next topic in 20 seconds 17:25:22 #topic this weeks FAH release 17:25:27 that's a correct summary 17:25:38 so we are delayed on this weeks release because of a bug in releng 17:25:43 images aren't building 17:25:45 :( 17:25:50 Might as well remove the meeting bug until the hw is available 17:25:59 https://pagure.io/atomic-wg/issue/273 17:27:01 so hopefully we'll get that fixed soon so image builds start working again 17:27:12 #topic open floor 17:27:17 actually 17:27:20 wait a sec 17:27:25 #topic kubernetes in f26 17:27:39 so.. we are getting down to crunch time for f26 17:27:54 so, the system container version jbrooks built works 17:28:01 who has had a chance to review jbrooks' work on containerized kube? 17:28:16 I have as well 17:28:25 works as far as I can tell 17:28:41 so how do we feel about this being mainstream for us? 17:28:55 One thing about containerized kube is that it works whether or not kube is in the image 17:29:06 jbrooks: right 17:29:07 I don't know enough about kube to really say 17:29:12 otherwise I'd look at it 17:29:21 I guess a big thing is -- how important is the size difference in the image to us? 17:29:25 so here's the steps required (1) need official builds of the contianers in FLIBS (2) need doc updated (3) need an official repo of the modified ansible install, or get Kubernetes to accept it 17:29:34 jbrooks: size difference? 17:29:42 i'm almost of the opinion that it would be good to have the f26 release to "trial" containerized kube to see if we want to remove it for f27 (in the fall) 17:29:43 The biggest downside to including it in the image is that you can't rpm-ostree install a different version 17:29:48 if we remove the packaged kube I guess 17:30:38 jbrooks: but there is only one rpm available in our repos at a given time 17:30:46 jbrooks: well technically if you can have other versioned images, updating the container is really easy 17:30:54 dustymabe, that's true 17:31:10 It's easy to use different versions w/ containers 17:31:24 how does everyone feel about having containerized kube as something we promote for f26, but also leave kube in the image 17:31:30 and then in f27 we can rip it 17:31:46 this is kinda how we *can* use overlay2 in f25 now 17:31:50 but it will be default for f26 17:32:05 its better than directly ripping it 17:32:11 or wait wasn't that the original plan 17:32:24 I'm looking at centos atomic vs centos atomic continuous -- the latter is missing kube, etcd, flannel, gluster and ceph, and is 100MB smaller, the qcow 17:32:25 yzhang: well right now in f26 it is *not* there 17:32:44 yeah all of those binaries are large 17:33:10 dustymabe, I'm +1 17:33:17 jbrooks, excellent 17:33:23 to promoting while we it still built in 17:33:27 everyone else? 17:33:27 dustymabe: that seems like a good plan. it means that we can get our users to test contianerized kube 17:33:33 to shake out issues 17:33:35 yes, and also give us feedback 17:33:36 dustymabe: seems like a good compromise +1 17:33:39 +1 17:33:45 walters: ? 17:33:55 jbrooks, we'll be addressing the replacement case at some point; that's https://github.com/projectatomic/rpm-ostree/issues/485 17:33:56 also, jbrooks++ 17:33:56 jberkus: Karma for jasonbrooks changed to 6 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:34:09 ... for figuring out a containerized kube that works even if you have kube installed 17:34:30 walters, cool 17:34:36 Who all are coming PyCon? 17:34:46 kushal: not the venue for that question 17:34:56 walters, and will users be able to install layered packages live in the f27 time frame? That'd make this even smoother 17:34:57 oops, pressed enter too fast 17:35:16 :) 17:35:37 jbrooks, we'll definitely have it under `ex livefs`, need more real world experience before saying how soon it could graduate from `ex` 17:35:47 ex == experimental 17:35:50 ? 17:35:52 yes 17:35:55 cool 17:35:58 ok moving to open floor 17:36:02 well, first 17:36:21 #agreed we will leave kube in the image for f26 and trial/promote containerized kube 17:36:30 #topic open floor 17:36:47 * roshi only had budget for one conference, so I don't get to go to pycon this year 17:36:51 will be at flock though 17:37:10 roshi: yay - /me hopes to go to flock 17:37:21 * sayan is going to PyCon 17:37:21 jdoss: did you have something for us? 17:37:29 but leave out gluster/ceph? 17:37:35 yea I do 17:37:47 This is in regards to Fedora Cloud image updates for EC2 and Vagrant. 17:37:55 Lack of updates on Fedora Cloud images can cause a poor end user experience. An example of this is with Fedora 25 Vagrant image it's out of date and if a user tries to create a vagrant file with an NFS mount due to this RPCbind bug https://bugzilla.redhat.com/show_bug.cgi?id=1401561. The bug has been fixed for sometime now but the current Fedora 25 base Vagrant box is out of date. My comany uses Fedora Cloud 25 for our developm 17:37:56 ent boxes and we have had to bake our own updated boxes (https://atlas.hashicorp.com/kennasecurity/boxes/25-cloud-base) to get things to work correctly. Most users won't have the time or skills to do this on their end. 17:38:47 Also, Fedora 26 Alpha Cloud images for testing are unavailable while Fedora 26 Atomic Alpha are. 17:39:02 jdoss: so we have a cloud SIG group for the cloud base image 17:39:16 I am willing to help make these updates and images. I just need to direction on that. 17:39:18 and we'd really like to get to a point where we are releasing images periodically for cloud base 17:39:36 maybe we can try to do this for f26 ? 17:39:50 jdoss: you are willing to help out with it ? 17:40:00 I am down to help out with that. Just having an updated image cuts my dev box build times down by like 10min. 17:40:07 sounds like jdoss can be a member of the cloud SIG 17:40:12 yep 17:40:21 sorry if this is the wrong meeting for bringing this up 17:40:43 jdoss: nope, you are fine. we don't have an official meeting for that group yet 17:40:46 https://pagure.io/cloud-sig 17:41:01 that is the pagure instance - i've been meaning to migrate some tickets from the atomic-wg tracker into the cloud-sig tracker 17:41:09 OK cool 17:41:40 #action dustymabe to migrate some cloud-base tickets to the cloud-sig tracker from the atomic-wg tracker 17:41:40 I just feel that if we make the cloud images easier to use in production and devel people will move to Fedora over using other distros. 17:41:57 :) 17:42:08 * dustymabe loves it when a new contributor starts using "we" 17:42:13 jdoss: welcome to the community 17:42:31 I offer myself as tribute. 17:42:44 haha 17:42:44 * jberkus gets out the silver knives and the incense 17:42:46 nice 17:42:54 anyone else have anything for open floor 17:43:07 My crying for the day is done. 17:43:19 jdoss: thanks for hanging around 17:43:24 Anytime. 17:43:32 ok we'll wind this down 17:43:33 welcome jdoss :) 17:43:42 Hi 17:43:46 It's me again 17:43:48 x3mboy: hi 17:43:50 welcome 17:44:02 Sorry, but Talking Points are not ready yet :( 17:44:27 dustymabe, thanks :D 17:45:04 jberkus: ^^ 17:45:29 hey 17:45:37 x3mboy: what time zone are you in? 17:49:13 did we lose him? 17:49:16 x3mboy: ^^ 17:49:26 UTC-3 17:49:28 My network disconneted. Did I miss open open floor? 17:49:40 deuscapturus: it's open floor now 17:49:46 deuscapturus: welcome back 17:50:00 excellent. thank you. I would like to follow up on https://bugzilla.redhat.com/show_bug.cgi?id=1423753 17:50:04 dustymabe, Sorry, I'm in several meetings 17:50:09 It's still the case that I cannot copy any AMI's.. old or new. 17:50:15 x3mboy: can we sync up on talking points tommorrow? work through them? 17:50:25 jberkus, of course!!! 17:50:36 deuscapturus: yep the issue is still open. sayan is very close to getting it fixed 17:50:43 thanks for pinging us on it again 17:50:52 great thanks! 17:50:52 Thanks, and sorry again, but is important to the marketing meeting 17:51:04 s/meeting/team/g 17:51:07 deuscapturus: do you have a specific AMI you are trying to access? Fedora Atomic or Fedora Cloud base? 17:51:19 x3mboy: that's fine, and it helps that you're here for reminders 17:51:33 x3mboy: jberkus schedule a meeting? 17:51:37 so that we don't drop it 17:51:57 dustymabe: yeah, bringing up a time zone converter 17:51:58 dustymabe: both. I've tried several. But it's the AMI generally in the 125523088429 account. 17:52:12 jberkus, dustymabe No problem from me 17:52:26 deuscapturus: yep. the reason I ask is because *new* images will have proper perms 17:52:35 x3mboy: I'm available 1800 UTC or later 17:52:42 Also, I would like to participate in this community more. Should I register with the chat bot? 17:52:42 we don't yet have a way to retroactively go back and fix old ones 17:53:04 deuscapturus: thats awesome, basically just hang out in #fedora-cloud 17:53:14 and come to these meetings 17:53:26 and volunteer when work exists that you want to work on 17:53:27 dustymabe: how does he register? with fedoraproject.org? 17:53:38 register for what exactly? 17:53:54 I notice that people .hello at the beginning of this meeting 17:54:26 deuscapturus: yep, do you have a FAS account? 17:54:31 with Fedora? 17:54:37 I think so... 17:54:40 .hello deuscapturus 17:54:41 deuscapturus: deuscapturus 'Theodore Cowan' 17:54:45 yep :) 17:54:46 there you go :) 17:55:09 #action x3mboy and jberkus to hold meeting tomorrow to discuss marketing talking points 17:55:17 ok I think we are done with the meeting 17:55:27 deuscapturus: let's chat in #fedora-cloud 17:55:32 thanks for coming everyone! 17:55:35 #endmeeting