17:00:12 #startmeeting fedora_atomic_wg 17:00:12 Meeting started Wed Jan 11 17:00:12 2017 UTC. The chair is jbrooks. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:12 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:12 The meeting name has been set to 'fedora_atomic_wg' 17:00:19 #topic roll call 17:00:25 .hello roshi 17:00:26 roshi: roshi 'Mike Ruckman' 17:00:28 .hellomynameis kushal 17:00:29 kushal: kushal 'Kushal Das' 17:00:32 .hello trishnag 17:00:33 trishnag: trishnag 'Trishna Guha' 17:00:36 .hello dustymabe 17:00:37 dustymabe: dustymabe 'Dusty Mabe' 17:00:38 .hello coremodule 17:00:40 coremodule: coremodule 'Geoffrey Marr' 17:00:49 coremodule: \o/ 17:01:03 .hello bowlofeggs 17:01:04 bowlofeggs: bowlofeggs 'Randy Barlow' 17:01:09 .hello sayanchowdhury 17:01:10 sayan: sayanchowdhury 'Sayan Chowdhury' 17:01:12 .hello maxamillion 17:01:13 bowlofeggs: maxamillion 'Adam Miller' 17:01:22 .hello maxamillion 17:01:23 * sayan is finishing off another meeting 17:01:23 maxamillion: maxamillion 'Adam Miller' 17:01:35 bowlofeggs: :) 17:01:51 #chair roshi kushal trishnag dustymabe coremodule bowlofeggs sayan maxamillion 17:01:51 Current chairs: bowlofeggs coremodule dustymabe jbrooks kushal maxamillion roshi sayan trishnag 17:02:09 #topic action items from last meeting 17:02:15 here now, sorry 17:02:19 .hello jberkus 17:02:20 jberkus: jberkus 'Josh Berkus' 17:02:30 #chair jberkus 17:02:30 Current chairs: bowlofeggs coremodule dustymabe jberkus jbrooks kushal maxamillion roshi sayan trishnag 17:02:37 * kushal has an open floor topic 17:02:54 jberkus to create tickets for prerequisites for removing kube packages from base ostree: https://pagure.io/atomic-wg/issues?status=Open&tags=remove-kube 17:02:59 ack, kushal 17:03:06 that's done 17:03:16 it seems like I'm forgetting some major prerequisite 17:03:23 That's the link I posted 17:03:51 I'm working on them -- the containers part is waiting on the release bits for the layered build service 17:04:31 I can build the master and node containers on which the individual components are based, but can't layer on them to build those components until they're releaseable 17:04:42 oh 17:04:49 testing, knew I was forgetting something 17:05:30 :D 17:05:37 next action, dustymabe and walters to discuss new release process 17:05:59 jbrooks: we've kind of already done that, no I just need to formalize and get with releng 17:06:05 s/no/now 17:06:12 I would say just drop it from action items 17:06:28 ok, added 17:06:36 OK, might be good to post something to the list 17:06:43 roshi: https://pagure.io/atomic-wg/issue/191 17:07:17 next action, dustymabe walters jberkus to explore overlayfs writeability issue 17:07:21 jbrooks: agree 17:07:35 some notes on overlayfs: https://pagure.io/atomic-wg/issue/186 17:08:16 AFAIK, that's done except discussing partitioning 17:08:19 * roshi takes a look 17:08:34 jberkus: right 17:08:36 jberkus, OK, carry it forward? 17:08:37 overlayfs may be *technically* not posix-compliant, but I can't get it to actually break anything 17:08:48 yes, we need to decide on the partitioning 17:09:17 #action dustymabe walters jberkus to discuss overlayfs partitioning issue 17:09:38 Anything to discuss on that right now? 17:10:07 jbrooks: I think it might be best if we leave it to the ticket 17:10:13 OK 17:10:20 take a look at the ticket. I think it's almost decided, though 17:10:28 dustymabe, next item is yours, blogging about overlayfs 17:10:58 jbrooks: so actually there is something I'd like to talk about there 17:11:07 Take it away 17:11:08 so there were some new features added to docker-storage-setup 17:11:32 I'd actually like to get those into fedora 25 before I blog, because I want users to be able to use the new DOCKER_ROOT_VOLUME=yes option 17:11:47 Is something in bodhi? 17:11:51 dwalsh has said that he is willing to merge the patches if we test it and tell him it's good 17:11:56 OK 17:12:10 Can you point us to a fork? 17:12:11 jbrooks: I'll create an ostree that people can rebase to to test it out 17:12:15 Cool 17:12:15 ok 17:12:22 are you guys willing to hammer on it a bit? 17:12:27 yes 17:12:34 the most important thing to to make sure behavior stays the same 17:12:42 given the same inputs as the past 17:12:43 dustymabe: do we have good docs on 'how to make and host your own custom ostress' ? blog post or otherwise? 17:13:06 maxamillion: good docs, probably not 17:13:15 there are lots of different ones 17:13:21 jason has done several blog posts and really it is in the fedora-atomic git repo 17:13:22 maxamillion: that's one of those things we *really* could use for our users 17:13:38 I've written a bunch -- I keep meaning to unify 17:13:43 maxamillion: these aren't good answers to your question 17:13:57 maxamillion, I don't think so :) 17:13:57 jberkus: +1 17:14:03 dustymabe: they are not :) 17:14:16 I'll update this one: http://www.projectatomic.io/docs/compose-your-own-tree/ 17:14:22 so "no" ... that answer is "no" ... which is fine, we should just look into making that a "yes" :) 17:14:23 I wrote that a while ago 17:14:29 jbrooks: +1 17:14:32 jbrooks++ 17:14:32 maxamillion: Karma for jasonbrooks changed to 3 (for the f25 release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:14:49 #action jbrooks to update compose your own doc 17:15:12 #action jbrooks dustymabe jberkus to collab on testing new docker-storage-setup patches 17:15:41 next action? 17:15:45 scollier to update ticket on Future of Fedora Dockerfiles 17:16:02 https://pagure.io/atomic-wg/issue/180 17:16:18 I don't think we have him atm 17:16:33 jberkus jbrooks to do initial draft of PRD and ping dusty and group for further review 17:16:59 on the topic of that ticket, I wrote a porposal for a workshop at DevConf (we would host it both in-person and virtually for those who can't attend) 17:17:05 so, we started on that 17:17:08 OK, Josh and I went through this, we made a bunch of edits, and I put in several comments: https://public.etherpad-mozilla.org/p/fedora_cloud_PRD 17:17:15 the former Cloud PRD is llllooooooonnnnnggggg 17:17:20 longer than the other PRDs 17:17:33 we're wondering if we can just cut a bunch of stuff 17:17:37 jberkus: did you ping us for feedback :) 17:17:38 we'd divvy up all the work into taiga cards (I'd pre-populate them when we get there) and then work on the Fedora-Dockerfiles at the workshop ... please check the mailing list and reply there 17:17:42 jberkus, yesterday I more or less got to the end, but I put in a bunch of "can we cut this?" comments 17:18:02 dustymabe, consider yourself pinged now 17:18:08 :) 17:18:51 maxamillion, could you add a link to that proposal in the issue? 17:18:59 jbrooks: will do 17:19:03 thanks 17:19:22 .hello scollier 17:19:23 scollier: scollier 'Scott Collier' 17:19:29 #chair scollier 17:19:29 Current chairs: bowlofeggs coremodule dustymabe jberkus jbrooks kushal maxamillion roshi sayan scollier trishnag 17:19:35 Hey Scott 17:19:43 heya jbrooks apologies. 17:19:56 no prob 17:20:28 So, on the PRD, if people could check it out and comment, that'd be great 17:21:17 jberkus: jbrooks: thank you for putting that together! 17:22:31 I don't know if it's legit to do: 17:22:48 #action all comment on PRD draft to discuss next week 17:23:37 The last action on the list is jberkus to lead planning some fedora-atomic work sessions around devconf 17:23:51 Any updates there, Josh? 17:24:08 maxamillion proposed something for FDLIBS for Friday afternoon 17:24:25 what other work sessions do we need 17:25:07 FOSP ? 17:25:19 we really haven't made much progress there from what I can tell 17:25:47 well, the folks who know how to do OS+Fedora seem to have no access to the HW 17:26:02 and if there's any real interest in Project FAO anymore, we would meet on that ... system containers and the like might be in a better place to actually get some action items lined up for it 17:26:31 there's still interest, we even put it in the draft PRD ;-) 17:26:42 These are session ideas for devconf? 17:27:11 are these sessions or just meetings that we can have "around" devconf? i.e. they don't have to happen on the actual days of devconf 17:28:12 yah, so the problem with meetings outside devconf 17:28:21 is that I'm not sure there are days that most of us are there 17:28:32 I'm there for 1 day before and 2 days after 17:28:48 * misc will be present 1 week before devconf but leave on monday 17:28:53 jberkus: +1 17:28:59 yeah I think the day before is the most likely day 17:29:09 there is also the need to get a room 17:29:23 I'm there 2 days before, zero days after and I already have a lot of meetings lined up for the 2 days before 17:29:42 I wish this was discussed before hand. Could have been a case for us to attend. 17:30:10 Yeah, I probably would have come 17:30:20 Count me in. :) 17:30:31 kushal: just like almost anything else, we waited til the end to plan things 17:30:38 there is an easy solution here 17:30:41 :) 17:30:50 kushal: it's going to be a 2 hour session and anything done in person can easily be done online ... so if we want to make this for after DevConf, we can setup a Virtual FAD if that is preferred 17:30:55 dustymabe: plan for devconf 2018 now ? 17:31:07 dustymabe, yup, and that also means we will miss. 17:31:26 maxamillion, not for this particular meeting, but if we have any other meeting etc as mentioned. 17:31:38 kushal: oh 17:31:40 fair 17:32:00 the only one I really know that has any agenda is that one that jberkus and I were talking about 17:32:07 maxamillion: yeah, virtual FAD would be useful once every two months or so I would say? 17:32:08 I can do a virtual FAD 17:32:17 I won't be at DevConf anyways :) 17:32:20 time zones might be a problem 17:32:26 was planning on being in the IRC though 17:32:34 hmmm 17:32:45 dustymabe: that's not a bad idea 17:32:49 given the number of people who *won't* be there, I'm dubious about doing anything around DevConf 17:32:57 Does someone want to take the ball on setting up a virtual FAD? 17:32:58 jberkus: true, unless we decided that we wanted to make that our #1 priority for that day 17:33:08 dustymabe: maybe you and I should just do an FAO hack session after devconf? 17:33:22 dustymabe: we can make a taiga kanban board for it, plan and groom before hand and then just have one day every couple months where we hack on things 17:33:36 jberkus: sure, just let me know what you are wanting out of it and I can try to make sure I deliver from a technical perspective 17:33:51 maxamillion: yeah 17:34:14 I don't want to overcommit to making it a bigger deal than it is 17:34:41 but i think just having us all together for a period of time periodically would make things start to "move" 17:34:43 dustymabe: +1 17:34:43 faster 17:34:50 yep 17:35:12 i'll open a ticket for us to discuss 17:35:17 Excellent 17:35:33 makes sense to me 17:35:38 dustymabe, is this discuss devconf or discuss virtual fads 17:35:55 virtual fads 17:35:59 #action dustymabe to start ticket to discuss virtual fads 17:36:04 on to tickets 17:36:15 isn't a "virtual fad" a meme? 17:36:34 ;-) 17:36:46 alright, new generic Atomic WG kanban board https://taiga.fedorainfracloud.org/project/maxamillion-fedora-atomic-wg/kanban 17:36:47 Ah, I haven't been breaking these action items into topics 17:36:55 #topic November 21 ISO is not bootable on UEFI 17:37:03 https://pagure.io/atomic-wg/issue/185 17:37:04 I'll reply to the mailing list with this discussion for posterity and I'll start putting cards in 17:37:35 I'll make an Epic for the Fedora-Dockerfiles stuff and I'll write up some docs on the process ... anyone have any preferences on where that stuff goes? 17:37:44 maxamillion: where's our kanban? 17:38:12 maxamillion: I created a ticket https://www.petfinder.com/petdetail/37080834 17:38:14 sigh 17:38:19 https://pagure.io/atomic-wg/issue/192 17:38:24 maxamillion, new wiki page linked to from the build service wiki page docs? 17:38:32 side note, we might adopt a dog 17:38:36 heh 17:38:58 hehe 17:39:01 maxamillion, btw, what is the full form of Epic? 17:39:06 a chihuahua, though? 17:39:09 heh 17:39:28 jberkus: https://taiga.fedorainfracloud.org/project/maxamillion-fedora-atomic-wg/kanban 17:39:32 sorry, I forgot to link it 17:39:33 jberkus: yeah, we already have one chihuahua, so... 17:39:42 also, I have no idea why maxamillion is in the URL ... taiga just does that 17:39:48 nice 17:39:50 I didn't slap my name in there ... I'd prefer it not be there 17:40:02 * jberkus needs to set up a personal kanban now that Atlassan bought Trello 17:40:21 ok, I guess we should talk about the topic 17:40:31 any updates on that? 17:40:40 or other dog options? 17:40:46 ;) 17:40:51 so.. the story is that there is some bug in pungi blah blah.. I'm still working with releng on it 17:40:55 jberkus: taiga.fedoraproject.org is open for business :) 17:41:17 OK 17:41:43 We already talked about #180 17:41:50 and #170 17:42:13 I guess I'll update #170 w/ call for comments 17:42:31 #topic Ship fedora-motd in F24 atomic image 17:42:36 https://pagure.io/atomic-wg/issue/160 17:42:40 Any updates on this one? 17:42:49 Is rtnpro here? 17:43:00 nope 17:43:13 ok 17:43:29 #topic Proposed website changes for Cloud Base → Atomic Host switch 17:43:34 https://pagure.io/atomic-wg/issue/158 17:43:49 This was last modified two months ago 17:44:21 that mentions a dustymabe meeting on this 17:44:40 so.. we haven't announced yet because there is an issue with the reported amis 17:44:43 but check it out 17:44:47 cloud.fedoraproject.org 17:45:07 Seems legit 17:45:08 :) 17:45:17 Maybe update that ticket? 17:45:23 i'll close this ticket 17:45:31 cool 17:45:36 and let the ticket in the websites repo serve for fixing the remaining issues 17:45:36 #topic design, deploy and document Fedora OpenShift Playground (FOSP) 17:45:54 misc, Do you have any news on this front? 17:46:11 jbrooks: nope, didn't I posted on the bug ? 17:46:22 right, sorry 17:46:26 https://pagure.io/atomic-wg/issue/153 17:46:27 OK 17:46:30 well, I do have news we are getting hardward 17:46:39 awesome 17:46:41 but for that, I need to decipher the catalog of super micro 17:46:58 that's like doing RSA computation but without a computer 17:47:17 I know you can do it :) 17:47:27 OK, let's move to 17:47:30 #topic open floor 17:47:37 kushal, you have an item? 17:47:44 Yes 17:48:10 Yesterday sayan and me proposed https://fedoraproject.org/wiki/Automated_AMI_test_and_release as F26 change. 17:48:22 What else have we put in from Atomic WG? 17:48:40 As changes? 17:49:06 kushal: have you sync'd with puiterwijk about autosigning that stuff? 17:49:27 jbrooks, Yes. 17:49:34 maxamillion, autosigning what? 17:49:40 kushal: also, is this something that could be incorporated with the releng-automation work? https://pagure.io/releng-automation 17:49:40 kushal, What's the deadline for this? 17:49:52 jbrooks, standard F26 change. 17:50:11 kushal: aren't there checksum manifests that are gpg signed? or some AMI record that's gpg signed? 17:50:21 maxamillion, I don't know any. 17:50:36 maxamillion, if there is any, we should sign those with help from puiterwijk 17:50:48 change due dates: https://fedoraproject.org/wiki/Releases/26/Schedule 17:50:59 kushal: alright, just checking ... for some reason I thought there was something to establish a chain of trust for the AMI images 17:51:02 maybe not 17:51:54 maxamillion, and for the second query, this is done on the fedimg side, I don't know how that fits in the automated release tools story. it should, but just don't know the steps/details. 17:52:04 maxamillion: also, is it possible to know when the compose id changes? 17:52:12 Correct. 17:52:26 Fedora-Docker changed to Fedora_Docker and autocloud and fedimg broke 17:52:27 Sudden change in the compose id breaks the other systems. 17:52:53 kushal: yes, dgilmore made that change I believe 17:53:25 dustymabe, that is okay, just that we should have been informed to update autocloud and fedimg. 17:53:32 dustymabe, as that breaks 2WA etc. 17:54:02 kushal: sure, I agree, I'm just telling you where to go have a conversation about it 17:54:03 Yes, I asked dgilmore later. At the first go I thought it was a typo 17:54:03 :) 17:54:14 :) 17:54:18 maxamillion: no current chain of trust to AMIs exist, no. 17:54:23 dustymabe, I know, just saying here. 17:54:32 sayan: so when can we expect to get those things fixed? 17:54:40 i know you were working on it earlier 17:54:53 we have a release that we will need to do early next week 17:55:16 puiterwijk: rocking 17:55:27 dustymabe: yes, that's the thing I'll do after the meeting 17:55:28 For future, I also want to remove the local qcow2 image testing, and then test only on the Fedora Infra Cloud. 17:55:46 dustymabe, https://bodhi.fedoraproject.org/updates/FEDORA-2017-6ae65d21bd 17:55:55 sayan: know when the compose id changes? you mean like when a compose is done? or something else? 17:56:06 We could have current stable kube in it if we test/karma 17:56:28 maxamillion: so the compose id used to be Fedora-Docker-* which changed to Fedora_Docker-* 17:56:32 kushal: why is that? didn't the local testing find a bug recently that wasn't showing up in Fedora Infra Cloud? 17:56:36 jbrooks: that should be in the updates-testing tree then 17:56:49 so.. should be pretty easy to get a system with it on there 17:56:50 Oh, we have that tree now? 17:56:57 jbrooks: we "should" 17:56:59 I've been meaning to ask you 17:57:03 sweet 17:57:08 sayan: that should all be in fedmsg/datagrepper just like it always has been, if it isn't then please let me know 17:57:57 maxamillion: dgilmore changed it to underscore, and would be changing the same for atomic too 17:57:58 maxamillion, It did, then we should mention running qcow2 images locally also as an option to people. 17:57:58 kushal: it does not break 2wa 17:58:04 I mean we should market more. 17:58:07 kushal: its a docker base image only compose 17:58:11 https://kojipkgs.fedoraproject.org/atomic/25/refs/heads/fedora-atomic/25/x86_64/testing/ 17:58:15 looks like it 17:58:33 Fedora-Atomic would become Fedora_Atomic 17:58:52 sayan: I guess I don't understand the question 17:58:55 dgilmore, ^^^ 17:58:57 sayan: what is it that you're trying to find out? 18:00:16 maxamillion: the value of the compose id in the fedmsg message changed 18:00:26 sayan: ok 18:00:38 the format changed so autocloud and fedimg started throwing errors 18:00:47 sayan: why does the format matter? 18:01:01 we're at the top of the hour, should we adjourn to #fedora-cloud 18:01:13 +1 18:01:17 #endmeeting