19:01:50 #startmeeting Fedora Cloud Working Group 19:01:50 Meeting started Wed May 13 19:01:50 2015 UTC. The chair is jzb. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:01:50 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:02:05 should we start pinging all members of Cloud WG before each meeting? 19:02:06 #chair jbrooks dustymabe 19:02:06 Current chairs: dustymabe jbrooks jzb 19:02:24 dustymabe: perhaps 19:02:44 https://fedoraproject.org/wiki/Cloud/Governance 19:02:48 #chair oddshocks 19:02:48 Current chairs: dustymabe jbrooks jzb oddshocks 19:03:04 .hellomynameis dustymabe 19:03:05 dustymabe: dustymabe 'Dusty Mabe' 19:03:12 number80: ping 19:03:12 jzb: Ping with data, please: https://fedoraproject.org/wiki/No_naked_pings 19:04:10 that needs to happen in every channel ever and private MSGs... 19:04:19 roshi: are you about? 19:04:28 I don't see scollier 19:04:34 * jzb dimly recalls something about PTO 19:04:41 let's get started 19:04:49 I just pinged theem in fedora-cloud 19:04:54 dustymabe: thanks 19:04:56 they can join in if they make it 19:05:25 we only have two tickets marked "meeting" atm 19:05:36 both of those are long-ongoing things about Docker images. 19:06:01 yep. we talked about Fedora-dockerfiles work last week 19:06:34 Yeah, I had an action item to see if docker hub would accept branches for automated builds. And it does: https://registry.hub.docker.com/u/jasonbrooks/fedora-dockerfiles/builds_history/202211/ 19:06:57 That build is pointed at a branch in my fork of the fedora-dockerfiles repo 19:07:48 jbrooks: sweet 19:08:16 so. should we concentrate on any items we need to take care of before F22 release 19:08:34 I know walters made a change for the UID/GID stuff. Do we have something that we can test yet? 19:08:43 jbrooks: so we should be able to have builds based on branches for all the dockerfiles/ 19:08:46 er, ? 19:08:57 jzb, yep, that ought to work 19:09:25 dustymabe: I think he sent something to the list about that. 19:10:02 * walters is here now 19:10:24 walters: excellent 19:10:31 walters: we were just talking about the UID/GID status 19:10:41 can someone pastebin the backscroll? (irc =( ) 19:10:47 walters: yes, just a sec 19:11:11 * mattdm mumbles about silly -1 19:11:22 walters: http://fpaste.org/221636/ 19:11:33 mattdm: hrm? 19:11:58 the images are here http://koji.fedoraproject.org/koji/tasks?state=all&owner=masher&view=flat&method=createImage&order=-id 19:12:06 jzb: (channel name. don't mind me) 19:12:26 mattdm: oh, I thought you meant somebody -1'ed something 19:12:33 jzb: nope sorry :) 19:12:42 +1s for everything today! 19:12:47 #info images for UID/GID change 19:12:54 #link http://koji.fedoraproject.org/koji/tasks?state=all&owner=masher&view=flat&method=createImage&order=-id 19:13:31 we'll need to add something to the release notes about the UID/GID change 19:13:35 http://koji.fedoraproject.org/koji/taskinfo?taskID=9716685 is a direct link that won't involve scraping the relational database 19:13:42 yeah 19:14:03 mattdm: because I'm lazy, is it too late to add changes to the release notes? 19:14:17 mattdm: I forget whether they're part of freeze or if changes are still accepted 19:14:24 walters: ok. will try to run that image through some paces today or tomorrow 19:14:29 dustymabe, thanks! 19:14:42 does anything need karma or are we past that? 19:14:43 the main ideal test would be rebasing to/from the centos downstream rebuild which has matching uids 19:15:12 walters: is the centos downstream atomic image released yet? 19:15:18 * dustymabe is out of the loop on that one 19:15:23 walters: do we have a CentOS build with matching UIDs? 19:15:25 it's being worked on 19:15:27 jzb they're part of freeze but changes might be accepted 19:15:50 there's a holdup around docker not building 19:16:06 #action jzb send patch for release notes about UID/GID change, verify text with walters for sanity before sending. 19:16:30 walters: do we need some sort of announcement as well? 19:16:40 AIUI this breaks rebasing from F21 to F22? 19:16:46 right 19:16:54 we have apparently been in the string freeze since *February* 19:17:02 here is the exception policy https://fedoraproject.org/wiki/Software_String_Freeze_Policy 19:17:03 Might be worth a Fed Mag post? 19:17:05 mattdm: that's... conservative. 19:17:12 * mattdm makes note to review this 19:17:16 mattdm: wait, software string policy? 19:17:19 because that ain't pragmatic 19:17:21 mattdm: that includes release notes? 19:18:16 jzb: apparently. see http://fedoraproject.org/wiki/L10N_Freezes#What_changes_are_affected_by_the_string_freeze.3F 19:18:46 #action mattdm to look at string freeze as it relates to release notes 19:19:01 that's crazypants 19:19:43 jzb: doesn't jibe with https://fedoraproject.org/wiki/Release_notes_process#The_Process 19:19:48 walters: sorry, what was that about Docker not building? 19:20:03 see point #13 there 19:20:16 the most recent docker from RHEL Extras apparently doesn't build with the latest golang 19:20:40 mattdm: so I need a patch to fedora-release-notes ASAP 19:20:49 from Optional; the CentOS Core SIG was going to work around it by pulling it the golang from CBS i heard 19:23:01 OK, anything else on the UID/GID change? 19:23:32 i think we're good.. i'll try to find centos atomic and rebase 19:23:40 i guess I could use RHELAH ? 19:23:50 walters: ^^? 19:24:13 it's on my todo to write a script which helps out with the f21 -> f22 rebase for simple cases, basically chowning your ssh key 19:24:15 hey dustymabe, sorry i'm late. 19:25:00 scollier: welcome! 19:26:31 as far as karma, https://admin.fedoraproject.org/updates/FEDORA-2015-8109/man-db-2.7.1-4.fc22 needs some 19:26:45 that will in turn let me build nfs-utils again for that bug which i linked to on the list 19:27:15 walters.. noted 19:27:24 will try to take a look at the same time 19:28:43 walters: http://fpaste.org/221643/ 19:28:58 walters: that's a quick quick quick rough cut at text for the release notes. 19:30:03 jzb, i'd say "A script to rebase between F21 and F22 is in development; check back (somewhere) for updates"? 19:30:04 any other topics this week? 19:30:22 walters: I'm *assuming* we'll have that before final? 19:30:26 not that I know of 19:30:40 btw I know think a few people hit https://bugzilla.redhat.com/show_bug.cgi?id=1216151 19:30:42 during test day 19:31:41 * oddshocks pops in late 19:32:03 dustymabe: do we know if that's a bug introduced somewhere in our chain or an upstream docker bug? 19:32:16 I don't know if anybody is working on it or not.. I think it is filed upstream as well 19:32:30 https://github.com/docker/docker/issues/12811 19:32:35 but rhatdan asked them to open a bz 19:33:24 so may be Fedora specific? 19:33:29 yeah 19:33:31 bummer 19:33:50 OK 19:34:06 we should probably have someone follow up on it to find out who needs to fix 19:34:29 #topic Talking Points for F22 19:34:35 quick note 19:34:47 please take a look at https://fedoraproject.org/wiki/Fedora_22_talking_points 19:35:04 and add things to the cloud section if you can think of anything we should be specifically calling out 19:35:30 we are very close to final release, we should probably have two or three articles for Fedora Magazine or other publications to highlight features in the cloud edition. 19:35:44 any volunteers? :-) 19:36:20 walters: I know you are overburdened with free time... 19:36:32 walters: but any chance of getting a list of improvements in rpm-ostree that affect F22? 19:36:42 it might be a cool article just to talk about the fact that you can rebase between centos and Fedora atomic 19:36:56 yeah, i would like to talk about that 19:36:57 except for the fact that CentOS isn't out yet 19:37:20 walters: if you want to send me a list or outline I can help flesh it out 19:37:31 walters: or if you want to write it, even better, but I know you're probably pressed for time. 19:38:53 i'm editing the wiki now 19:38:59 walters: excellent 19:39:02 I would *really* love to see at least one non-atomic talking point 19:39:24 #action jzb follow up on rpm-ostree article w/walters 19:39:28 mattdm: vagrant boxes? 19:39:32 Since people didn't seem ready to make atomic the main offering, we should have *something* to say about that main offering 19:39:34 oh, wait, we don't get those :-( 19:39:39 jzb: did that get unbl.... 19:39:40 jzb: do we have non-atomic vagrant boxes? 19:39:42 jzb-- 19:39:44 jzb-- 19:39:46 jzb-- 19:39:52 mattdm: ? 19:39:54 * mattdm shakes fist 19:39:57 haha 19:40:01 you got my hopes up there :) 19:40:21 mattdm: well, we can talk about improvements in Docker, the Docker image, and uh... 19:40:37 kushal could put out another talk about tunir maybe? 19:40:49 I'm open to suggestions 19:40:57 we broke the image down a bit (by splitting out kernel package) ? 19:40:57 "higher level of automated testing"? 19:41:09 that's not the shiniest, but maybe the best we've got? 19:41:18 dustymabe, didn't that split happen in 21 19:41:30 jbrooks: perhaps... hmm 19:41:39 mattdm: do we have that? 19:41:42 I thought he sent out an email recently where he highlighted that 19:41:55 yes that was 21 19:42:25 I think we're also _growing_ in size overall. TC1 image is 218M 19:42:33 uh, but there are no TC2 or TC3 images? 19:42:44 * mattdm pardon me if that's known 19:43:00 mattdm: yeah I know at least roshi knows about it. I assume rel-eng is well aware 19:43:05 oddshocks: might know more? 19:43:06 dustymabe: k 19:43:10 dustymabe: do we have a ticket? 19:43:17 but anyway we can't lean on "it's even smaller!" :( 19:43:38 mattdm: true 19:43:52 I think for now that's the best stuff we have.. mostly highlighting atomic 19:43:55 and tunir 19:44:06 oh 19:44:10 that reminds me. 19:44:11 Atomic has been where a lot of effort has been focused 19:44:51 we need an entry for the Docker bas eimage 19:44:53 *base image 19:45:24 walters: that's a good talking point yes. and cloud is as good a place for it as any 19:45:33 dgilmore shot down moving it to Cloud but then things are weird because Base isn't part of the website... 19:45:54 at the moment it ends up as its own toplevel https://dl.fedoraproject.org/pub/alt/stage/22_TC3/Docker/ 19:45:58 walters: from the pov of the talking points it can go in cloud 19:46:08 yeah 19:46:14 and we need to revisit for F23 19:46:16 pretty much mostly releng cares about what directory it is in on the mirrors 19:46:24 we shot down moving to cloud for F22, I don't think that's permanent 19:46:37 nothing is permanent :) 19:46:43 and I don't care what directory it goes in - I care about getting it more attention on the release pages. 19:46:50 speaking of docker base image, is lsm5 (or someone) lined up to push it to the docker hub? 19:46:51 mattdm: so true, so true 19:47:07 #info philosophy 101 in here today 19:47:21 mattdm: is that a rel-eng task or specifically an lsm5 task? 19:47:37 mattdm, we really need to bring the lsm5 -> docker hub thing into daylight. Right now it's totally opaque to me, and dgilmore also mentioned fedora rel-eng doesn't do it or know about it 19:47:47 jzb: specifically a lsm5 task because the process is ugly and manual and requires github 19:47:52 he's cc'd in https://fedorahosted.org/rel-eng/ticket/6170 19:48:00 walters: tahnks 19:48:05 although that's f21 19:48:13 yeah 19:48:31 #action mattdm to ping lsm5 about f22 upload readiness 19:48:34 mattdm: who would be responsible for making that less ugly and opaque? 19:48:39 jzb: Docker Inc 19:48:51 action: mattmd to fix Docker 19:48:56 lol 19:49:09 no but i'm emailing lokesh right now :) 19:49:14 mattdm: thanks sir 19:49:37 mattdm: maybe we need something by way of a release checklist for Cloud specifically? 19:51:11 jzb: yes that's a good idea 19:51:25 let's through kushal under that bus since he's not here. all in favor? :) 19:52:15 +1 19:52:17 :) 19:52:24 +1 19:52:35 #action kushal create release checklist for Cloud 19:52:46 #action jzb apologize to kushal for assigning tasks when he's not here. 19:52:52 lol 19:52:59 a'ight 19:53:09 * mattdm s/through/throw/ geez 19:53:19 we have 7 minutes remaining in the hour... any other (quick) topics? 19:54:00 +1 for setting fuse 19:55:06 boom 19:55:10 #endmeeting