20:11:52 #startmeeting 20:11:52 Meeting started Fri Feb 11 20:11:52 2011 UTC. The chair is dgilmore. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:11:52 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:12:01 #meetingname releng 20:12:01 The meeting name has been set to 'releng' 20:12:07 #topic init 20:12:41 notting: Oxf13: skvidal: abadger1999: nirik: ping 20:12:45 yo 20:12:58 * abadger1999 around 20:13:02 idle. 20:13:05 * nirik is here. 20:13:16 should we invite the new guys? 20:13:22 notting: yep 20:13:42 nb: ping 20:14:10 * jsmith lurks 20:15:09 #info present jsmith notting nirik abadger1999 dgilmore Oxf13 20:15:18 #topic mass rebuild 20:15:40 so other than being crammed it went reasonably smoothly 20:15:59 there is still some packages needing building 20:16:02 sounds like there might be some corner case tagging issues? 20:16:11 possibly that also 20:16:33 im waiting on just under 5k rpms to be signed 20:16:45 then i can do a branched compose and see how we are looking 20:16:55 rawhide finished today? 20:17:18 good question 20:17:19 looks like it did. 20:17:20 it started :) 20:17:24 but hasn't synced yet? 20:17:49 seems to be not finished 20:17:59 oof 20:18:01 it's still going 20:18:59 presumably, all the rebuild-induced deltas 20:19:13 probably yeah 20:19:23 its still running mash 20:19:51 well, hopefully it will be along... 20:20:18 once the packages are signed ill manually run a branched compose 20:20:32 ok. 20:20:48 try and get the tc2 images created 20:20:51 how do we want to handle updates? I didn't push any this morning... 20:21:14 anything else on the rebuild anyone wants to add? 20:21:39 #topic updates 20:21:43 ok updates 20:22:16 id like to wait till i get the f15 packages all signed 20:22:23 then do updates pushes 20:23:10 ok. You want me to push them all, or should we split up the pushing? 20:23:26 i need to grant access to the fedora-15 key to at least nirik 20:23:40 with branched we just push testing right? 20:23:50 regular is nightly ? 20:23:54 no,it's more complicated than that 20:24:04 * nirik tries to recall. 20:24:06 nirik: well we push all except for when frozen 20:24:23 there's also stable pushes, for things that pass testing. it just tags them into dist-f15 20:24:31 when we are in the feature freeze we manually include the builds that fix the blockers 20:24:36 and push testing 20:24:56 notting: right thats a per build thing 20:25:18 once pushed to stable it goes to dist-f15 20:25:32 we need to make sure that the packages are signed 20:25:40 and have signed copies written to disk 20:25:48 then the branched compose will pull them in 20:26:08 there is 2 phases of pushes 20:26:20 the one issue is that because stable-bound things are moved from updates-testing to dist-f15, once you push stable updates, you need to wait for the next branched compose before you push testing again 20:26:21 when we are not frozen we can push to stable and testing 20:26:26 otherwise packages disapper 20:26:31 disappear, even 20:26:36 when frozen we can push to testing and selected builds to stable 20:26:47 yeah, thats the wrinkle I was trying to remember. 20:27:26 notting: right 20:27:33 so push to testing 20:27:39 then do the stable push 20:27:51 then after branched can be done again 20:28:25 or perhaps do testing in the morning, then wait in case we want more testing pushes during the day, and do a stable later in the day when things are quieted down? 20:28:36 or that 20:28:42 that can work too, if you're willing to keep up with it 20:29:03 if we don't do a stable push, the branched compose will pick those up, right? 20:29:10 so, perhaps no stable pushes unless requested? 20:29:22 or it only picks up things that have been pushed already to stable? 20:29:33 nirik: we have to do the stable push for things to hit the branched compose 20:29:42 ok. 20:30:20 we should update http://fedoraproject.org/wiki/Pushing_fedora_updates 20:30:46 yeah 20:30:50 ill do that 20:30:51 we just be careful with what we push during freezes 20:30:59 according to the wiki, we're not frozen now 20:31:13 yeah. 20:31:44 #action dgilmore to update wiki with info on pushing for branched 20:31:47 so, I am happy just adding F15-testing to the morning pushes I have been doing... perhaps then someone else could do the stable ones later in the day? 20:32:02 notting: right next tuesday we hit freeze 20:33:51 http://poelstra.fedorapeople.org/schedules/f-15/f-15-releng-tasks.html lists Feb 15 20:34:09 #link http://poelstra.fedorapeople.org/schedules/f-15/f-15-releng-tasks.html releng tasks 20:34:38 or if folks want I can do stable too, but then we couldn't do a later one... 20:35:25 nirik: maybe we should make the cutoff 5pm est 20:35:33 sounds reasonable. 20:35:34 do a stable push then 20:35:43 nirik: are you doing pushes via web or cli? 20:35:49 cli 20:36:14 and we can do testing pushes during the day 20:36:16 ok. may need to fix the cli to be able to only do a testing/stable push 20:36:24 notting: ive never used the web to do a push 20:36:25 I think it can do that... 20:36:34 notting: it can do it 20:37:07 bodhi -P --push-release=F15 --push-type=testing 20:37:11 yeah. 20:37:19 that works. i think it was --push-type=stable that didn't 20:37:27 (unless it got fixed) 20:37:32 it was push-type=testing that didn't work, and I fixed it 20:37:45 thanks lmacken. ;) 20:37:53 notting: yeah stable builds during frozen we cant use --push-type=stable 20:37:59 fixed in 0.7.10 20:38:11 we need to push only approved blocker builds 20:39:22 * nirik nods. 20:39:24 ok i think we have a plan for updates 20:39:40 #topic alpha tasks 20:39:56 i think were on track 20:40:06 the mass rebuild put us a little behind 20:40:22 but not to the extent that its going to cause delays 20:40:56 we have a bunch of stuff that we need to do over the next few weeks 20:41:14 I guess we still need to block / retire orphans? 20:41:21 nirik: working on that 20:41:28 cool. 20:41:39 * nirik wonders if he can add one of his... 20:41:47 nirik: you can 20:42:01 the first RC for alpha is due next thursday 20:42:46 please let me know if i missed something 20:42:56 notting: they going to be blocked today? 20:43:15 dgilmore: wanted to give a little more warning 20:43:24 notting: ok no problems 20:44:17 once i have a branched compose im going to do tc2 compose. would have been good to have them blocked 20:44:55 oh well 20:45:23 let's re-run the script and see how bad it would be 20:45:30 notting: ok 20:45:55 #topic koji admins 20:46:22 so i noticed yesterday that caillion tagged stuff into dist-f15-override 20:46:26 which suprissed me 20:46:53 i revoked his admin privlidges and gave hime fedora-override 20:47:11 those are separate? 20:47:16 but it made me think we should check who has admin privlidges and if they really need it 20:47:24 notting: they are as of last week 20:47:46 rather than grant admin to tomspur and nb i did the same as was done for epel 20:48:05 there is a epel-override and fedora-override groups 20:48:28 to tag into overrides you only need to be a member of the appropriate groups 20:49:20 my main concern was that things were tagged into overrides and not being documented 20:49:26 maybe that doesnt matter 20:49:29 who are the current set of admins? 20:50:05 notting: let me look in the db 20:50:08 probably needs to be 'anyone who pushes updates' (on top of existing koji guys like mbonnet/mikem/etc) 20:50:19 yeah, the command line doesn't seem to have a way to list them out. 20:52:14 anyhow, we could do that out of band? or do we want to do it now? 20:53:24 lets do it out of band 20:53:45 what are peoples thoughts 20:53:55 wfm 20:53:59 sure. 20:54:06 * skvidal is sorry he wasn't hear earlier 20:54:48 sounds good. 20:56:56 * nirik has a topic if we still have time/for open floor. 20:57:34 * notting has one too 20:58:12 nirik: its yours 20:58:20 #admin nirik notting 20:58:29 #admins nirik notting 20:58:45 you want #chair 20:58:48 ? 20:58:55 #chair nirik notting 20:58:55 Current chairs: dgilmore nirik notting 20:58:59 nirik: yep 20:59:03 the floor is yours 20:59:17 #topic Feedback on multidvd 20:59:21 https://fedorahosted.org/rel-eng/ticket/4342 20:59:28 we were asked for feedback on this issue a while back. 20:59:33 we should really respond. ;) 20:59:46 nirik: i spoke with spot on this yesterday. he was going to come up with some saner tooling 21:00:06 we can put it on alt 21:00:16 excellent. So, it might end up being something we produce? or at least host? 21:00:30 but my biggest concern is all the fighting over the tooling 21:00:35 yeah. 21:00:48 I can see this being a lot more useful for ambassadors than most users. 21:01:00 nirik: well the board wants releng to be the producers. but i feel it has no place as part of the regular release 21:01:11 so alt seems like a good place for it 21:01:40 sure, but that still leaves who produces it and with what tools... 21:01:45 but it sounds like thats being worked on. 21:01:53 nirik: well releng produces 21:02:00 ok. 21:02:06 its the tools that are in question 21:02:21 cool. So, if you or spot could update the ticket with some info that would be great. 21:02:30 sure will do 21:02:32 thats all I had. Just wanted to bring that up 21:02:38 nirik: thanks 21:02:43 notting: floor is yours 21:02:46 dgilmore: From my perspective, I don't think it needs to be tied to the release schedule 21:03:05 dgilmore: But I would like it produced by rel-eng, with tools that are in Fedora (for obvious reasons) 21:03:07 ok, just wanted to comment on rawhide this week 21:03:23 jsmith: considering we dont make livecds with updates it makes sense to do it with the rest of the release 21:03:33 it failed twice due to a very bizarre error copying the files. no obvious reason it would have happened other than 'something else running in parallel', which didn't seem to be the case 21:03:35 #topic rawhide this week 21:03:41 we couldn't reproduce it 21:03:46 and it passed that part today 21:04:16 #info rawhide failed twice this week, in nonreproducible fashion 21:04:29 #info it passed this area today 21:04:36 #info watch in case it happens agian 21:04:39 #undo 21:04:39 Removing item from minutes: 21:04:43 #info watch in case it happens again 21:04:57 just stating that for the record. not sure what else to do at this point 21:05:12 notting: i guess all we can do is keep an eye on it 21:05:14 yeah, it was a weird one. 21:06:20 alright anything else? 21:06:57 * nirik has nada 21:07:00 not i 21:07:03 if not ill close up in 30 21:07:50 #endmeeting