15:01:37 #startmeeting Fedora Base Design Working Group (2015-02-20) 15:01:38 Meeting started Fri Feb 20 15:01:37 2015 UTC. The chair is haraldh. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:38 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01:54 * jreznik is here 15:02:16 ping dgilmore masta vpavlin msekleta 15:02:29 hello everyone! 15:03:59 #chair haraldh msekleta jreznik 15:03:59 Current chairs: haraldh jreznik msekleta 15:04:01 hmm 15:04:32 lets wait a couple of minutes for the others 15:04:37 hi haraldh 15:04:47 #chair haraldh msekleta jreznik dgilmore 15:04:47 Current chairs: dgilmore haraldh jreznik msekleta 15:04:57 Hi 15:05:40 #chair haraldh msekleta jreznik dgilmore vpavlin 15:05:40 Current chairs: dgilmore haraldh jreznik msekleta vpavlin 15:06:07 #meetingname Fedora Base Design Working Group 15:06:07 The meeting name has been set to 'fedora_base_design_working_group' 15:06:32 #topic New chairman of the Base Design Working Group 15:07:07 As you might know, Phil will leave Red Hat in the next couple of days. 15:07:22 So, we will have to find a replacement for him. 15:08:03 I volunteer 15:08:28 anyone else? 15:09:25 I will send out an email about this, so we can vote in the next meeting. 15:09:27 haraldh: I would if i knew I could dedicate the time it deserves 15:09:36 haraldh: happy to have you lead things 15:10:18 we also want a totally new member then 15:10:28 we will 15:10:28 "Each voting member of the working group will confirm their continued 15:10:29 membership every six months. In the event that a current voting member 15:10:29 relinquishes their seat, the remaining voting working group members are 15:10:29 responsible for appointing a new voting member to fill the seat from the 15:10:29 active Fedora Base community via majority consensus." 15:12:11 I figure we should announce on fedora-devel that we are looking for new member, right? 15:12:26 correct 15:12:47 well, maybe it's time to let my place too, I'm not that good member (as base shifted a bit) 15:14:41 can you elaborate about the "shift" ? 15:21:20 haraldh: well, initial idea before we established wg was, we are going to have out of sync releases, where base would coordinate it 15:21:56 we're not doing it, so my taskjuggler foo (scheduling tool) is not that important and that was why phil selected me 15:22:30 #action haraldh will write an email to fedora-devel that we are looking for new members. 15:23:16 jreznik, maybe we should revive this discussion again 15:23:37 or maybe it's still too early 15:23:47 that, too 15:24:15 jreznik: we can't support different lifecycles without major change and a lot of automated tooling 15:24:34 #topic Open Floor 15:24:38 jreznik: but shipping different products at different points in time I would strongly opose 15:25:23 dgilmore: that's why I say it's too early to restart such discussion 15:25:33 and it would be just waste of time 15:25:59 jreznik: well I do not think we will ever be able to support shipping different products on different schedules 15:26:09 but we can end them on different days 15:26:16 with the right things set up 15:26:33 yep 15:27:24 but then again in base we only rely on the shortest lifecycle 15:28:10 so if workstation goes on a 4 month release cycle and server on 12 month, base is 4 month 15:28:49 haraldh: nope 15:29:28 haraldh: we would still need to ensure that their is bugfixes in base things until server is eol 15:29:49 there 15:30:36 btw. I hope workstation guys actually wants longer release cycle than shorter 15:30:50 haraldh: base feature development would be over when the releases ship 15:31:05 jreznik: ive heard talk of 7 months 15:31:19 just enoiugh time to switch over to the next release 15:31:47 where server is wanting 19-25 months 15:31:59 cloud I am not sure 15:32:20 dgilmore: cloud would be probably the fastest moving, they are pretty agile there 15:32:31 jreznik: right 15:32:46 for ws, I heard even more than 7 months as the plan is to have stable devel ws 15:33:12 you can never make people happy even if you would release more than one product with different cycles 15:33:29 i am sure 15:35:06 I still somehow like the idea of having too maintained releases but with different rules 15:35:09 for updates 15:36:11 that would be a lot of work. it would need a lot of tooling and process work for enforcement 15:36:17 fn as faster moving, fn-1 as slow moving... not "almost no updates allowed for both" 15:37:17 haraldh: 4 month release cycle ? 15:37:31 mclasen, just an example :) 15:42:54 anything else? 15:43:18 vpavlin: where is the docker base image at? 15:44:27 dgilmore: sorry, I am confused - what do you mean now? 15:44:46 vpavlin: what is the current status? 15:45:47 we have filed a PR on docker/library for new version of rawhide, but tianon is against my changes (which I did to make systemd work) 15:45:55 https://github.com/docker-library/official-images/pull/497#issuecomment-74945551 15:46:41 well..I also wanted to change f21 image Dockerfile a bit so that systemd works for both 15:47:26 dgilmore: Are image building successfully in Koji again? (I haven't checked that for few days..) 15:47:49 docker really needs to get an api to update the registry 15:48:03 vpavlin: last nights f22 base image built 15:48:29 rawhide is likely to fail for awhile 15:48:47 we are undergoing some pretty major changes in process for f23 15:49:04 ok 15:49:12 I'll chceck f22 then 15:50:53 I have tested docker on armv7hl 15:51:07 but we do not have tooling to make base images :( 15:52:13 vpavlin: for f22 will we just make the base image> 15:52:15 ? 15:52:24 or is there some extra images we want? 15:53:26 I think we can "release" the rest under Fedora-Dockerfiles for now 15:53:44 okay 15:54:09 anything we do not build as part of the Fedora compose process can not be called Fedora 15:54:18 but has to be labeled as a remix 15:54:38 including replacing fedora-release and fedora-logos 15:55:54 not sure if that matters much 15:56:08 I do not totally understand how Dockerfiles work 15:56:18 All other images should be based on the Fedora base image 15:56:42 sure, but if we do not make it, it can not be called Fedora 15:56:45 You specify base image which is basically a rootfs that is just extended 15:57:17 vpavlin: is it something users do? or is it done by someone in the middle? 15:59:01 dgilmore: both - we do that in https://github.com/fedora-cloud/Fedora-Dockerfiles, users can also do that 15:59:21 vpavlin: so users doing it for their own use is fine 15:59:34 vpavlin: but others doing it and calling it fedora is not 15:59:48 vpavlin: even if those others is you 15:59:57 hmm 16:00:09 vpavlin: if you do it, it has to be called a remix 16:00:33 dgilmore: Interesting 16:01:16 I think scollier is maintainer of that repo 16:01:20 vpavlin: https://fedoraproject.org/wiki/Legal:Trademark_guidelines?rd=Legal/TrademarkGuidelines#Copies_of_unmodified_Fedora_media 16:01:41 vpavlin: it comes down to trademark rights 16:02:07 vpavlin: it can only be called fedora if produced by fedora releng as part of the compose process 16:02:15 obvious way out: join the fedora rel-eng team... 16:02:45 mclasen: That probably wouldn't help - Fedora-Dockerfiles are built in Docker Hub 16:02:58 So not as part of the compose process 16:03:18 vpavlin: we can produce the images 16:04:53 vpavlin: anyway, I think we need to talk more about this, and possibly engage legal 16:05:02 vpavlin: we likely need to ask legal 16:05:12 they may turn around and say it is fine 16:05:35 so long as it is unmodified fedora provided software 16:06:14 which it is 16:06:39 but you do not get trademark use rights today 16:06:50 and you need to 16:07:09 the way to get the rights to the trademark today is to have it built as part of fedora 16:08:58 ok 16:11:05 vpavlin: can we setup a time to go over how it all works next week? 16:11:32 Sure, I'll be on pto thu and fri 16:12:09 okay. lets aim for early in the weke then 16:15:31 I am probably fine with any of Mon, Tue, Wed, but it would be good to have Scott there also as he maintains those Dockerfiles 16:15:42 okay 16:16:02 vpavlin: I will try set up a meeting for the three of us 16:16:18 thanks 16:20:19 anything else for the meeting? 16:22:19 #endmeeting