17:00:19 #startmeeting F21 Beta Go/No-Go meeting 17:00:19 Meeting started Thu Oct 23 17:00:19 2014 UTC. The chair is jreznik. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:19 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:20 #meetingname F21 Beta Go/No-Go meeting 17:00:20 The meeting name has been set to 'f21_beta_go/no-go_meeting' 17:00:41 Hi all! 17:00:41 morning 17:00:43 #topic Roll Call 17:00:57 who's around today? 17:00:59 .hello sgallagh 17:01:02 sgallagh: sgallagh 'Stephen Gallagher' 17:01:08 .hello kevin 17:01:08 .hello roshi 17:01:09 nirik: kevin 'Kevin Fenzi' 17:01:12 roshi: roshi 'Mike Ruckman' 17:01:13 .hello dmossor 17:01:16 danofsatx: dmossor 'Dan Mossor' 17:01:54 #chair nirik sgallagh roshi adamw 17:01:54 Current chairs: adamw jreznik nirik roshi sgallagh 17:02:38 let's wait a minute and we can start to finish earlier - a lot of stuff to be done, readiness meeting in two hours then... 17:03:00 * danofsatx goes to nuke his lunch 17:03:14 Nuke it from orbit. It's the only way to be sure. 17:03:28 * adamw is filing the compose request atm 17:04:01 Do we have a rel-eng representative? 17:04:25 dgilmore, are you around? good point sgallagh 17:05:05 #topic Purpose of this meeting 17:05:25 #info Purpose of this meeting is to see whether or not F21 Beta is ready for shipment, according to the release criteria. 17:05:26 #info This is determined in a few ways: 17:05:28 #info No remaining blocker bugs 17:05:29 #info Release candidate compose is available 17:05:31 #info Test matrices for Beta are fully completed 17:05:59 #topic Current status 17:06:47 #info no RC available for Fedora 21 Beta, RC1 request filed at the moment 17:06:54 so we have a problem with ARM 17:07:08 beta TC4 ARM images had the generic-release issue and no-one's fixed that that I can see 17:07:20 we can either pull in generic-release 21.7 to RC1, or do a quick fix to the kickstarts 17:07:27 i think we're also waiting for the workstation tree being dropped per fesco 17:07:38 i think changing fedora-arm-base.ks the same way someone changed fedora-live-base.ks would do the job 17:07:42 * nirik saw a commit to do that earlier. 17:07:53 and we have no cloud images for TC4 17:07:55 nirik: to do which? 17:08:13 adamw, wouldnt the overall fix, also fix the issue for arm? 17:08:34 nirik, ah, indeed 17:08:37 adamw: drop workstation tree 17:08:52 ok, should be gone in the next compose 17:09:05 https://lists.fedoraproject.org/pipermail/rel-eng/2014-October/018625.html 17:09:33 pwhalen: i would expect that generic-release 21.7 would avoid the problem, yeah, but we haven't really done a test compose with it included 17:09:42 pwhalen: so far the fix for other images was a change to fedora-live-base.ks 17:09:49 #info workstation tree dropped 17:09:51 but arm images don't use that kickstart, they start from fedora-arm-base.ks 17:11:12 I think it would be wiser to pull in generic-release 17:11:17 just adding @fedora-release-nonproduct to fedora-arm-base.ks should 'fix' it (or at least behave the same as the lives, we haven't done a full test compose to ensure everything works the way we expect yet) 17:11:18 if we really want to do anything with rc1 this week, less intrusive way is preferred I'd say 17:11:47 adamw: fedora-arm-base should support products though 17:11:56 it would be a bad idea to make that non-productized 17:12:00 adamw, i see. I pinged Peter but he isnt around right now. I can ask if can make that change before the compose, or dgilmore. 17:12:04 sgallagh: well, it's the same thing we did in fedora-live-base.ks 17:12:15 sgallagh: fedora-workstation-packages drops it again and pulls in the product group instead 17:12:25 fedora-live-base.ks is a bit different since it only affects Workstation; other products aren't live-based 17:12:28 i can change it if we decide it's the right thing to do, i have commit rights 17:12:39 cool, ok 17:12:47 sgallagh: is arm planned to be productized - or what products are currently supported on ARM? 17:13:19 kalev: afaics only fedora-arm-kde, fedora-arm-lxde, fedora-arm-mate, fedora-arm-minimal, fedora-arm-soas, fedora-arm-workstation and fedora-arm-xfce include fedora-arm-base.ks 17:13:26 kalev: unless something else is pulling it in some other way 17:13:46 ah, no arm-server or arm-cloud? 17:14:17 i don't think we build such things ATM, and if we did, i don't think they'd use fedora-arm-base.ks 17:14:19 scary to do these kinds of changes the very last minute, if we botch them we'll definitely miss this week's release 17:14:41 well, we have to do something or the ARM images will be generic-ized, which we don't want 17:14:52 i guess i'll just throw the change at fedora-arm-base.ks and we can cross our fingers 17:14:57 id be willing to throw another vm up to test arm if need be 17:15:02 Sorry, urgent phone call. Be back soon 17:15:03 kalev: I'm becoming more inclined to let everything settle down a bit... reading all this last minutes stuff 17:15:38 it has become a case of what is solid and ready hasnt it ? 17:15:54 i sent the change, it's done. let's see what happens. :P 17:16:25 :))) 17:16:50 ok, so if there anything now we have to do for RC try? 17:17:03 adamw, thanks 17:17:04 how is arm composed, does it use pungi and friends? 17:17:29 jreznik: no, i just need to finish the compose request 17:17:53 is it the same system (pungi / anaconda) as is used for DVD and netinstall generation? 17:18:01 adamw: ok 17:18:25 so today we're definitely no-go, let's plan the next steps 17:18:41 I expect everyone agrees with todays no-go 17:18:56 yeah 17:19:00 for the record, on what criteria? 17:19:10 we have no rc 17:19:11 yes, no-go today, based on no RC being available 17:19:12 danofsatx: no release candidate available 17:19:21 rgr 17:19:40 I'm back. Reading scrollback now 17:20:08 #agreed Fedora 21 is No-Go due to no release candidate available as of 2014-10-23 17:20:28 #topic Beta release plan 17:20:48 we have the option to repeat Go/No-Go meeting tomorrow 17:21:13 a) we need RC, b) we need QA coverage and c) bits has to be available for releng tomorrow 17:21:58 looking on the whole effort put into it today, I think we want to try it but waiting for opinions 17:22:15 kalev, appliance-tools is used for arm images 17:22:37 I think this is primarily up to QA: are you folks willing and able to validate the RC compose? 17:22:58 going to be hard for arm validation, looking at 8 hours for the compose to land.. 17:23:00 sgallagh: yeah, definitely 17:23:28 pwhalen: which arm image do you guys consider release blocking? 17:23:35 minimal and kde 17:23:59 do we want to pull the SDDM CVE fix into RC1? 17:24:15 https://bugzilla.redhat.com/show_bug.cgi?id=1149610 17:24:21 what's the severity? 17:24:25 local privesc 17:24:36 impact=important 17:24:47 requires a spin-kickstarts change, unfortunately 17:26:02 what's the minimal impact in criteria? 17:26:19 jreznik: it's an FE not blocker 17:26:29 just wondering if it's important enough that we should pull it 17:26:34 I'd be inclined to say no if we are planning a marathon test session... the less changes the better 17:26:35 aha, sorry, I saw it but... 17:26:54 guess people can get the fix with an immediate update 17:27:03 there's no impact on live since it has no account security 17:27:09 adamw: well we said from some rh clasification it's blocker, I'm just not sure from which level :) 17:27:30 final IIRC 17:27:48 https://fedoraproject.org/wiki/Fedora_21_Final_Release_Criteria#Security_bugs 17:27:50 so it's a final blocker 17:27:55 ok 17:28:13 OK, RC1 request filed 17:28:24 without sddm, right? 17:28:40 https://fedorahosted.org/rel-eng/ticket/6010#comment:11 17:29:11 ok 17:29:39 and so there's still the question opened - Go/No-Go 2nd try tomorrow the same time? 17:30:38 jreznik: +1 17:30:45 +1 17:30:49 maybe we should ask releng folks first what time we'd need to know the Go / No-Go tomorrow so that they wouldnt' have to work over the weekend 17:30:54 adamw, roshi - it's really up to you what QA can make 17:31:02 we can try. 17:31:08 the test load is rather heavier in 21 than previously. 17:31:26 and of course, other teams are supposed to be pitching in 17:31:38 server folks should be testing server (sgallagh does, thanks sgallagh), workstation folks testing workstation, etc etc 17:31:44 kalev: it's still in the middle of day in the us 17:31:49 hey, we now have a thing which makes it pretty easy to see the test load: 17:31:53 https://fedoraproject.org/wiki/Test_Results:Fedora_21_Beta_TC4_Summary?rd=Test_Results:Current_Summary 17:32:00 love the summary 17:32:17 everything listed as Alpha or Beta there (except Security Lab, forget Security Lab) needs running 17:33:00 btw, for people who hate editing wiki text to enter results, i wrote a stupid thing that lets you submit results from a console app: https://www.happyassassin.net/wikitcms/ 17:33:04 I can try for an all nighter once the compose lands 17:33:07 kparal asked in the office for help, so maybe in the morning, if RC would be usable, we can ping all guys who volunteered to help 17:34:38 dgilmore: what's the latest time we have to say go on Friday? 17:36:09 but seems like we have agreement to try it, let's make the timing with Dennis later - I'll send announcement to the list after readiness meeting 17:36:22 for now I'd say the same time as today, may change 17:37:07 sure 17:38:15 ok 17:38:41 #info we will try Go/No-Go #2 tomorrow 2014-10-24, the same time (17:00 UTC) but may change, same channel - jreznik to ask dgilmore and announce it later today 17:38:57 #topic Open floor 17:39:17 btw. thanks for help everyone! 17:40:46 anything else today? 17:41:17 * roshi has nothing 17:42:33 ok, so thanks again, I'll end the meeting soon 17:43:08 just a reminder - readiness meeting later today19:00 utc 17:43:15 #endmeeting