19:00:01 #startmeeting EPEL (2016-03-09) 19:00:01 Meeting started Wed Mar 23 19:00:01 2016 UTC. The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:01 Useful Commands: #action #agreed #halp #info #idea #link #topic. 19:00:01 The meeting name has been set to 'epel_(2016-03-09)' 19:00:01 #meetingname EPEL 19:00:01 The meeting name has been set to 'epel' 19:00:01 #chair smooge nirik Evolution bstinson avij 19:00:01 #topic aloha 19:00:01 Current chairs: Evolution avij bstinson nirik smooge 19:00:11 hello world 19:00:20 hey all 19:00:33 morning 19:00:39 hello 19:00:59 hey. 19:01:03 hello all 19:01:19 looks like we have a full board today 19:01:34 I don't have much in the agenda as I have been up to my neck in something else 19:01:44 * nirik has been busy also. 19:01:48 #topic outstanding issues 19:02:16 ok any outstanding issues needing to be looked at? 19:02:28 Django14 should be back in 19:02:51 I haven't seen the emails about orphans lately.. I am checking to see if I am spam blocking them for some reason 19:03:01 has anyone else seen them? 19:03:05 the centos folks (Arrfab) were interested in an update or something of a guess at 'official' epel builders for armhfp 19:03:07 I don't think till has had a chance to run them 19:04:35 dunno. It would take figuring how to point that arch only to centos armv7 and then we could do bringup like we did for ppc64le 19:05:34 is there a released centos7 for armv7? 19:05:35 would this be within the same koji setup (thus possible nvr issues?) 19:05:57 I don't think we can do it in the same koji setup according to Dennis and Peter 19:06:16 Evolution: yes, the way ppc64le was done was to enable it in koji, then do scratch builds of everything, then import just the ppc64le builds and logs in 19:06:18 nirik: yes, it's released. Primarily it's 'userspace' driven due to the idiotic kernel fragmentation 19:06:43 but that's an armv7 thing 19:06:59 smooge: I thought dgilmore said we could... 19:07:57 I think that was where we had our breakdown in communication 3-4 meetings ago. He said it could but it would take a bunch of code changes and no one had looked at doing that. 19:08:55 You and I asked for more info on what those needed to be.. but he got run over by the F24 atomic bus 19:08:56 what is the proposal? 19:09:21 dgilmore: build armv7 epel against centos... 19:09:28 and the rest against rhel 19:09:35 thats doable 19:09:47 is that just different external repo urls? 19:09:50 but there will be issues at rhel point releases 19:10:01 we just add extra external repos yeah 19:10:31 there is a few ways epel for armv7 could be done 19:10:31 or for that matter actually, it could be done on kojipkgs? 19:10:34 dgilmore: that won't cause issues with nvr collision for noarch? 19:10:35 all have a cost 19:10:47 Evolution: no why would it? 19:11:05 Evolution: koji tracks the external repos things come from 19:11:13 noarch rpms would come from rhel 19:11:21 which may be interesting 19:11:33 ie for http://kojipkgs.fedoraproject.org/repo/rhel/rhel7/$arch/rhel-7-server-rpms/ where $arch is armv7 redirect to a centos one... 19:11:34 but should be fine 19:12:02 noarch subpackages in RHEL wouldn't cause issues with that? 19:12:07 nirik: yeah. we will likely have to add a empty rhel repo 19:12:16 and add empty placeholders for CentOS 19:14:45 dgilmore: how would we go about implementing and testing this? 19:14:56 Evolution: stage koji 19:15:04 yeah, we have a staging koji 19:15:30 Evolution: the tricker part is how we deal with things when rhel 7.3 comes out and CentOS 7.2 is the latest 19:15:34 so who can do this? is it something that bstinson and I can do or does it have to be fedora engineering? 19:15:48 I could giveyou permission to do it 19:16:02 would just need koji admin in stg right? 19:16:16 Evolution: main things, we need a copy of CentOS for armv7hl 19:16:21 then point it all at it 19:16:26 nirik: yep 19:16:27 yeah, we were planning on mirroring it... 19:16:36 smooge was going to look into that I think? 19:18:12 is all of this driven in ansible now? i'm sure we can help submit patches on the infra side 19:18:34 actually this is all in the koji db as far as I know... 19:18:43 its all in the koji db 19:18:49 we just need teh right commands 19:19:08 setting up the mirroring would be in ansible 19:19:11 well, the mirrors i mean 19:19:13 yeah 19:19:33 yeah. Should be pretty close network wise... 19:19:44 I am having network issues 19:20:41 freenode having doom. ;( 19:20:45 smooge: .. some network issues here as well, but back again 19:20:48 ah ok 19:21:00 so my last comment was 19:21:02 there was a 4 minute lag or so 19:21:02 Evolution, what does userspace mean.. everything but the kernel is provided? 19:21:19 smooge: correct. 19:22:08 Evolution, to mirror it do you guys use a fedmsg report or just ok for me to rsync 2x a day? 19:22:48 rsync should be fine. I *think* bstinson was looking at fedmsg for the future. 19:22:52 bstinson: thoughts? 19:23:09 fedmsg on our end is a ways away yet 19:23:13 bstinson, dgilmore I just need to put in a cron job in batcaves ansible to do the mirror 19:23:34 it should be done by tomorrowe 19:24:16 smooge: ping me if you need anything, you can either grab the whole tree or just mirror.c.o/altarch/ 19:24:17 smooge: did we want to put it in /pub? or just somewhere koji can get it? 19:24:24 nirik, dgilmore, I could mirror it on ftp_fedora :) 19:24:41 nirik, where does koji need it to be? 19:24:45 well, my worry there is that people will start using it as a centos mirror and add to our load 19:24:55 nirik, me too. 19:25:27 nirik, my oriignal plan was based off of pre-rsync doom 19:25:28 * dgilmore needs to run 19:25:39 see you dgilmore .. sorry for dropping off mid sentence 19:25:57 well, we still have rhel7beta and rc on /mnt/fedora_koji/rhel/ so we could put this on /mnt/fedora_koji/centos/ 19:26:18 ok can do that. 19:26:21 anyhow, we can sort that out of meeting 19:26:44 Evolution, I will have it mirrored somewhere by end of week. Next week we can work on the koji parts 19:26:54 sounds good. 19:27:02 and I need a rasp 2 or 3 19:27:18 which will be my next thing to deal with 19:27:33 bstinson, Evolution were there any other issues. 19:27:57 s/issues./issues?/ 19:28:32 not currently. i686 will be a thing as well, but I think one at a time is probably best. 19:28:39 yeah 19:28:43 I will mirror both at the same time 19:28:50 they don't take up much disk space 19:28:58 compared to our normal mirrors 19:29:20 #topic rechartering EPEL 19:29:53 so a couple of Fedora groups got rechartered last week. I hope to spend some time looking at their charters on Sunday 19:30:19 so I can write something up and have y'all critique it next meeting 19:31:46 +1 19:31:58 sorry I think I am getting lagged and probably kicked off again 19:32:18 I am going to move to openfloor and if I do people can close if no one has anything 19:32:24 #topic open floor 19:32:36 anyone have any issues.. if they are still on IRC :)? 19:33:01 nope. everything is awesome. :) 19:33:03 yselkowitz, you had something a couple of meetings ago. I don't know if we ever got to it. 19:33:41 smooge, I don't have anything right now 19:34:11 when we get closer to aarch64 enablement then I will 19:34:30 aarch64 will be fantastic to have. 19:35:25 you know that people will be all like "Oh cool, but will it run on my Pi3?" 19:35:48 * smooge laughs like it is funny but cries inside. 19:35:48 no, it won't :-) 19:36:12 tibbs, did you have anything you needed for us to address? 19:36:13 "just as soon as the pidiots add uefi support and sane upstream kernel support 19:36:28 Evolution, that is slated for the pi5 19:36:34 smooge: Don't think so. epel-rpm-macros is in the EPEL5 buildroot now. 19:36:37 yeah right... 19:36:42 yay! tibbs++ 19:36:42 smooge: Karma for tibbs changed to 7 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 19:36:44 So far I haven't noticed anything breaking. 19:36:54 o_O 19:37:01 tibbs|w: someone wanted the %epel macro from epel6... 19:37:08 ok final we are 1 year away from retiring EL5 19:37:15 Feel free to file a bugzilla ticket. 19:37:23 https://bugzilla.redhat.com/show_bug.cgi?id=1320497 19:37:51 Oh, that's against epel-release so I would never have seen it. 19:38:14 oh right. we should move it over 19:38:34 Well, either that or just do it in epel-release as is done in el7. Unless that's not actually it's done. 19:38:48 * smooge assigns epel-release to tibbs 19:38:51 (I haven't looked and all of my test EL vms are down currently.) 19:39:56 Well, what's the best place for them? 19:40:21 It's kind of weird to have epel-release containing macros, and also have epel-rpm-macros. 19:40:24 I don't think it matters much. 19:40:35 In el6, there's macros.epel. 19:40:35 probibly should move all the macros to the macros package 19:40:41 Sorry, in el7, there's macros.epel 19:40:46 yes, it predates the macros package 19:40:50 In el6, there's.... 19:40:58 macros.ghc-srpm. ??? 19:41:02 yep 19:41:19 Is epel-release in the epel buildroots? 19:41:22 yes 19:41:36 I think we should move the macros out of it tho... 19:41:43 put them all in one place 19:41:44 So it really doesn't matter, but I think it would make sense for it not to have macros. 19:41:45 I am +1 on that 19:41:49 right 19:42:30 Last to epel-release-5 was done by Seth in 2010. Sadface. 19:42:37 :( 19:43:11 :( 19:43:21 Is there any specific issues or policy involved with changing the epel-release package? 19:43:30 nope, shouldn't be. 19:43:46 there's a link that points to the newest one, but I think thats automatic updated 19:43:53 correct 19:43:58 OK, if someone wants to change the component on that ticket then I'll have a look at cleaning things up. 19:44:06 sure, thanks. 19:44:08 Kind of buried at the moment. 19:44:14 thanks tibbs 19:44:18 That seems to be a continual state these days. 19:44:19 No problem. 19:44:29 ok I have one last itme 19:44:49 did we get a go for F24alpha next week? 19:45:21 And thanks for pinging me; I wouldn't have noticed the meeting otherwise. 19:45:28 no problem tibbs 19:46:10 smooge: yes, alpha is go 19:46:15 thanks 19:46:37 #info expect rsyncs of EPEL to be slower due to alpha and netapp maintenance. 19:47:06 Any other open flood stuff ? 19:47:25 if not closing in ~2 minutes 19:50:14 #endmeeting