16:35:59 #startmeeting RELENG (2015-03-30) 16:35:59 Meeting started Mon Mar 30 16:35:59 2015 UTC. The chair is dgilmore. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:35:59 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:36:05 #meetingname releng 16:36:05 The meeting name has been set to 'releng' 16:36:05 #chair dgilmore nirik tyll sharkcz bochecha masta pbrobinson pingou 16:36:05 Current chairs: bochecha dgilmore masta nirik pbrobinson pingou sharkcz tyll 16:36:05 #topic init process 16:36:28 * sharkcz is here 16:36:38 * nirik is here. 16:37:05 * threebean is here 16:38:05 actually, gotta take my partner to work real quick. will brb. 16:38:57 #topic #5967 Enable source repo generation 16:39:03 lets get moving 16:39:08 https://fedorahosted.org/rel-eng/ticket/5967 16:39:26 so this is something I just plain do not want to enable 16:39:36 the cost of teh extra arch is high 16:39:46 and it does not give any benefit 16:40:18 can anyone think of a compelling reason to enable it 16:41:05 I think there is enough possibility of variance that we gain nothing 16:41:26 does anyone have any opinion? 16:42:41 well, this is for Koschei right? 16:43:36 * nirik looks for the ticket link 16:44:58 so, this would make srpms for every arch? or ? 16:45:50 mizdebsk: you around? 16:46:33 I'd like to hear some of the other possible uses mentioned... not sure I can think of many/any. ;) 16:48:05 so, I'd say punt this until mizdebsk can be around, and/or ask for more uses in ticket? 16:49:14 hum... 16:49:51 I'm not sure I understand the impact, but it would eat up storage space, and time is spent generating repos. Right? 16:50:16 it would make srpm repos for every repo 16:50:37 it would add an extra arh to the repos being generated 16:50:42 for every single buildroot 16:51:16 nirik: it has a high cost and we can not selectively enable it 16:51:32 yeah, so I guess we could just say, sorry, not at this time... 16:51:44 we have a lot of buildroots/targets these days. ;( 16:52:59 yep 16:53:04 the request seems to be mostly about depsolving, so I wonder if there is another way around 16:53:36 masta: right, and the issue is that depsolving from the srpms is wrong 16:53:54 because it does not deal with things like arch specific BuildRequires 16:54:17 the only way to be 100% sure is to rebuild the srpms for every arch and make sure 16:54:44 hrm... I can see that, yes 16:54:50 somewhat related is https://fedorahosted.org/fpc/ticket/303 16:55:22 I will close the ticket 16:55:35 #topic #5886 need method for distributing urgent fixes... urgently 16:55:42 https://fedorahosted.org/rel-eng/ticket/5886 16:56:48 so, how about I just make a proposal in ticket and we can discuss it next week and pick it apart? 16:57:01 because there's a bunch of tradeoffs. 16:57:02 nirik: sure 16:59:02 we basically just need to pick a set of things and then try and get it implemented. 16:59:06 (before the next urgent thing) 16:59:36 right 16:59:59 probably should consult the security response team also 17:00:12 yeah 17:00:44 #topic #6110 allow sudo for sysadmin-releng on sign bridges 17:00:50 https://fedorahosted.org/rel-eng/ticket/6110 17:00:55 nirik: is this done? 17:01:12 was I supposed to do it? 17:01:41 I think I was. ;) let me look. 17:01:47 okay 17:01:51 lets just do it 17:02:01 #info dgilmore or nirik to get it done today 17:02:04 #topic Secondary Architectures updates 17:02:05 #topic Secondary Architectures update - ppc 17:02:17 sharkcz: how is ppc? 17:02:44 we declared our Alpha RC6 as GO today, so we look good 17:02:51 cool 17:03:10 #info F22 alpha for ppc is go 17:03:15 dgilmore: note that the secondary bridge is not in ansible, etc. I need to redo them, but haven't gotten a window when they were not in use. ;( 17:03:25 nirik: okay 17:03:27 #topic Secondary Architectures update - s390 17:03:44 sharkcz: how is s390? 17:04:26 s390 will start with Beta publicly, but no issues from the builds point view 17:04:41 okay 17:05:05 #topic Secondary Architectures update - arm 17:05:14 http://dl.fedoraproject.org/pub/fedora-secondary/releases/test/22_Alpha/Server/aarch64/ 17:05:27 there is a Alpha compose in place for aarch64 also 17:05:37 #info aarch64 Alpha shipping 17:05:48 #topic Open Floor 17:05:57 anyone have anything for open floor? 17:06:16 I had one thing... let me see. 17:06:41 there's some folks seeing weird output from ppc koji? https://github.com/fedora-infra/fmn/issues/83 17:06:47 sorry, arm koji 17:07:04 and https://github.com/fedora-infra/fmn/issues/71 17:07:25 not sure if they mean much, but we should fix them if we can... 17:08:06 I can also paste them if folks don't want to look at github, just a sec. 17:08:34 nirik: well all secondary arches are implicated 17:08:37 http://paste.fedoraproject.org/204835/14277353/ 17:08:48 not all secondaries are running the latest koji code 17:08:53 yeah, not sure if it's a plugin issue then? 17:09:09 yeah. which we need to fix as soon as we can get time to. ;) 17:09:28 koji itself is fine 17:09:38 it would need looking at what fmn is doing 17:10:03 it could be that we changed koji behaviour but the secondaries have an old koji with old behaviour 17:10:13 though we try and avoid breaking things like that 17:10:26 fmn does this: https://github.com/fedora-infra/fedmsg_meta_fedora_infrastructure/blob/develop/fedmsg_meta_fedora_infrastructure/buildsys.py#L60-L158 17:10:49 to construct the body of its emails 17:11:04 file_base = 'https://kojipkgs.fedoraproject.org/work/' is wrong for secondary arches 17:11:29 and /work should not be used in urls 17:11:30 oh yeah, it's armpkgs / ppcpkgs ? 17:11:45 armpkgs s390pkgs 17:11:50 yes. (that doesn't explain the tracebacks though 17:11:51 not sure there is a ppcpkgs 17:12:27 it is not something we are foing to fix here 17:12:29 going to 17:12:40 hum... I'm guessing somehwere in python needs to have int() wrap the code inserting to the postgre db ? 17:13:05 will need to debug the fmn code 17:13:24 ok, just wanted to bring it up. 17:13:37 we don't have to fix it now, just making folks aware it needs to be looked at .;) 17:13:52 other news: I updated the releng-dash 17:13:52 needs fixing, but also seems like a trivial bug to fix. 17:13:55 https://apps.fedoraproject.org/releng-dash/ 17:13:59 can we move the bug to somewhere that is free and open 17:14:20 it now gets the branched/active status from pkgdb so it should need to be patched every 6 months as the release cycle unfolds. 17:14:27 s/should/shouldn't/ 17:14:46 docker and vagrant images are there now too. 17:14:59 okay cool 17:15:42 lmacken pointed out this morning that it doesn't show the new atomic repo status. but that's the last open ticket I have for it https://github.com/fedora-infra/fedora-releng-dash/issues/17 17:15:44 I want to propose to adjust the meeting time not to DST 17:15:56 I have one thing - seems I have a fix for missing Server group in anaconda in recent composes (at least Beta TC5) - because pungi drops the server environment because it removes newly introduced server-hardware-support group in comps - https://git.fedorahosted.org/cgit/spin-kickstarts.git/commit/?id=21545e84d68155d724cae73b689721c260937a1f 17:15:56 s/not/now 17:18:43 probably we should tie the meeting time to eg. EST zone, not UTC 17:19:20 Quick note: Bodhi will start composing f21-updates-testing atomic repos with the next push. 17:21:01 threebean: cool cheers 17:21:07 tyll_: change it to what? 17:21:51 dgilmore: to 15:30 UTC 17:22:56 or keep it at 11:30 EST/EDT (the US eastern tz) 17:23:10 * nirik doesn't care. ;) 17:23:25 * masta is indifferent 17:23:27 sharkcz: I do not like that idea 17:23:53 I prefer to deal with meetings in UTC or CST/CDT 17:24:07 I always miss meetings if teh time is in US Eastern 17:24:21 just because I forget and mess it up 17:24:31 then 10:30 CST/CDT 17:25:14 * dgilmore forgets if we have historically chnaged with DST 17:25:27 I am not really fussed either way 17:25:32 last year we did 17:25:44 when US DST started no one wanted it to follow 17:25:55 I did ask in a meeting 17:26:07 it was because it did not work for me 17:26:31 now that Europe changed, the time is less convenient for me 17:26:48 same here 17:26:51 the current time is less convenient 17:27:58 well that is regrettable. 17:28:05 tyll_: can you start a thread on the mailing list please 17:28:18 we can discuss and work something out 17:28:30 ok 17:29:01 I guess there will be no meeting next monday? 17:29:10 tyll_ & sharkcz what TZ offset are you two? 17:29:17 tyll_: there will be a meeting 17:29:27 tyll_: it is a normal work day 17:29:29 masta: UTC+2 now 17:29:39 dgilmore: ah, interesting, here it isn't 17:29:39 masta: CET/CEST same as tyll_ 17:30:22 tyll_: the US does not have any gazetted days off 17:30:30 you get what the company offers 17:30:39 I see 17:31:21 what is happening on Monday? 17:31:32 masta: easter Monday 17:31:35 * masta is jealous 17:31:42 at least in Australia it is a public holiday 17:31:54 we get Friday off work 17:32:18 we (in CZ) do work on Friday, but have Monday off 17:32:42 I'm that guy that always shows up to the office, wondering why it's empty, nobody there. Holidays! 17:32:56 australia has Friday and Monday off 17:33:07 yeah, friday is "spring holiday" :) 17:33:35 great, so one half the team is on holiday Friday, the other half on Monday... 17:34:11 anyways... 17:34:14 so the plan is to have a meeting on Monday 17:34:21 anything else? 17:35:10 * masta heads off 17:35:13 thanks dgilmore 17:36:22 * tyll_ too 17:36:23 thank you all 17:36:27 #endmeeting