15:00:47 #startmeeting Fedora QA meeting 15:00:47 Meeting started Mon Jun 27 15:00:47 2016 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:47 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:47 The meeting name has been set to 'fedora_qa_meeting' 15:00:53 #meetingname fedora-qa 15:00:53 The meeting name has been set to 'fedora-qa' 15:00:59 #topic Roll call 15:01:04 morning folks, who's around for the QA meeting? 15:01:12 * cmurf is in the ether 15:01:12 * pschindl is here 15:01:13 * satellit listening 15:01:17 * garretraziel is almost present, but not really, just pretending 15:01:18 * tflink is here 15:01:22 * sumantro is here 15:01:28 * brunowolff is here 15:01:32 garretraziel: you're finally catching on :) 15:02:36 I have to adjust to proper timezone, from UTC+2 to CEST :-) 15:03:09 you should make your meeting bot less honest, garret 15:03:45 alrighty, thanks for coming, everyone 15:03:48 garretraziel.honesty-- 15:03:50 * garretraziel is here 15:03:58 #topic Previous meeting follow-up 15:04:00 garretraziel: better 15:04:57 * coremodule is here. 15:05:12 #info "adamw and kparal to come up with list of 'blocking' bugs for graphical system upgrade" - we did that, they know what needs fixin', we're still waiting for the new builds to be added to the update 15:05:32 so that's in hand 15:05:51 #info "adamw to sync with pjones on grub2 blocker and try to get an RC built ASAP" - well, that's old, of course I did it and we got an RC2... 15:06:03 * pjones did it too 15:06:25 #info "adamw to try and figure out a way to get the FCoE test run" - tried, failed, story of my life. software emulated FCoE is hard. 15:06:29 * adamw hands pjones a cookie 15:06:40 mmm, cookies. 15:06:49 * kparal is late 15:06:52 don't suppose anyone has an FCoE setup in their garage they're hiding> 15:07:02 * adamw throws a spiky cookie at kparal 15:07:33 the reallocation failed grub bug seems to affect quite a lot of systems 15:07:51 I'm worried about that 15:08:15 yeah, me a bit too 15:08:21 but we're on previous meeting follow-up atm 15:08:46 "coremodule and kushal to try and test+karma https://admin.fedoraproject.org/updates/FEDORA-2016-714119426f" - hmm, lemme check into this one 15:09:01 ah, the cloud-utils bug. 15:09:10 #info "coremodule and kushal to try and test+karma https://admin.fedoraproject.org/updates/FEDORA-2016-714119426f" - they did that, it went stable 15:09:19 * coremodule tested it, worked good! 15:09:21 allllrighty, that's all the action items from last time - was there anything else to follow up on? 15:09:23 thanks coremodule! 15:09:39 adamw, No problem! 15:10:44 okey dokey, moving on then 15:10:48 #topic Fedora 24 retrospective 15:11:36 so, looking back on f24 as a product and the qa process...any concerns, thoughts about what was good/bad, ideas etc? 15:12:06 I like the change in how the kickstarts repo works. 15:12:55 livemedia-creator should have better documentation but works 15:12:56 The random build failures were very frustrating. 15:13:05 +1 15:13:20 satellit: since you've got it to work, maybe you could help with writing a wiki-style guide? 15:13:39 I have it on my wiki...f24 page 15:13:53 https://fedoraproject.org/wiki/How_to_create_and_use_a_Live_CD still has instructions for livecd-creator, so we could do with updating it 15:14:14 https://wiki.sugarlabs.org/go/Fedora_24#livemedia-creator 15:14:15 satellit: do you have a link? 15:14:36 #info satellit suggests we need better documentation for livemedia-creator (the newer live image compose tool now being used in production) 15:14:56 f23 does not seem to work except for url or nts? 15:15:00 This was my first time working with fedora qa team and I really loved the whole team and qa process :) 15:15:13 #info satellit has written up steps for livemedia-creator at https://wiki.sugarlabs.org/go/Fedora_24#livemedia-creator 15:15:38 I got it to work and once I figured out that I needed the no virt option and avoided copying rpms that I had locally it worked well. I used that to test the bios clock fix. 15:15:40 sumantro: glad to hear it :) thanks for jumping in 15:15:49 .fas linuxmodder 15:15:49 linuxmodder: linuxmodder 'Corey W Sheldon' 15:15:53 super late sorry 15:16:00 brunowolff: any chance you and satellit could work together to update the fedora wiki page? 15:16:05 * adamw trying not to take all the action items 15:16:30 if not I can help tag team on the wiki 15:16:48 * satellit glad to help on wiki page 15:17:00 I think we need a more appropriate blocker bug badgering system than adamw having to go poke people with sticks every week or two. 15:17:04 I can help with the wiki :) 15:17:12 ok, that sounds like enough cooks :) 15:17:24 I will double check the command line I used for the simple no mock or vm case and make sure that is available on the wiki. 15:17:48 #action satellit, brunowolff, linuxmodder and sumantro to update https://fedoraproject.org/wiki/How_to_create_and_use_a_Live_CD with livemedia-creator instructions 15:18:05 cmurf: what were you thinking of? 15:18:08 ack 15:18:39 maybe a coded in inactivity timeout 15:19:08 adamw: I don't have any terribly diplomatic ideas 15:19:10 i think there's a lot of value in manual pokes, even if there's an automated system' 15:19:16 but adding something automatic couldn't hurt 15:19:16 that scrapes vacation cal and if not on cal then auto send a reminder email to fas or other eamil 15:19:19 (much) 15:19:40 i have one or two terribly undiplomatic ideas 15:19:44 let's hear em! 15:19:52 or do it simpliar to pesign its auto but has a manual check 15:20:05 undiplomatic eh? 15:20:13 no that's not a good idea for Monday morning, besides doesn't zodbot have a censor? 15:21:04 er, what? 15:21:12 ^^ 15:22:14 #info cmurf suggests improving the way we poke maintainers about unfixed blocker bugs 15:22:15 The question really needs to be posed to people who have blocker bugs how they want to be poked with a stick, because already there's an update for a bug in the bug report which generates an email. 15:22:36 * adamw can't parse that last bit 15:22:43 "there's an update for a bug in the bug report which generates an email." ? 15:23:33 Yeah that's ugly. Anytime there's an update added to the bug report, it causes an email to go out to everyone cc'd on that bug. 15:24:20 So that's a kind of automated notification. But I think the issue is email sucks as a way of urgently notifying people of things. 15:25:03 well, i do ping people on IRC sometimes 15:25:19 not sure what other choices we have 15:25:22 OK that involves you, and hence the baby sitting problem. 15:25:43 more bugzilla mail is, imho, not a route that has ever led to more communication. 15:25:50 Exactly 15:26:14 So maybe zodbot could start poking people with IM's? 15:26:49 Or track them down on Hangouts or whatever they use... 15:26:52 I think a certain amount of baby sitting is really what we need. People will generally respond better to that than automated notifications. 15:27:12 so what some sort of parse of outstanding blocker bugs and ping $fas_nick " you paln to update / deal with this? " 15:27:38 cmurf, ^ 15:28:05 right 15:28:25 i guess i'd be worried that it'd nark people off 15:28:33 FAS containing a "preferred means of urgent notifications" might be useful if we end up having more than one supportable way of pinging people 15:28:40 adamw: umm, fucking good? 15:29:01 i mean more in the way of 'i'm blocking this damn bot and not fixing the bug just because' than 'hey whoops i'd better fix that!' 15:29:15 It's better for other people to get narked off because of 4 slips? 15:29:36 would this system have avoided any of the f24 slips? 15:29:42 adamw: ok well if they do that, they're pissing in the swimming pool and on themselves. 15:29:53 i mean, take the final slip - it's not like pjones didn't *know* there was a blocker bug to fix. he did. but he also had a bunch of rhel deadlines. 15:29:56 adamw: possibly 15:30:04 Punishment is not a good way to motivate people. They might avoid by fixing the bug, but they can also avoid it by stopping work on Fedora entirely. 15:30:17 the problem there is that we need three more pjoneses (what a terrifying notion), not that we need a more irritated pjones (also a terrifying notion) 15:30:20 * satellit afk 15:30:46 adamw: this is an on-going issue not just about f24 15:31:10 or just few bugs. It's always the case there are blockers that just don't get on someone's radar early enough for whatever reason. 15:31:35 i guess what i'm saying is just that i can't really think of many cases off the top of my head where we would've avoided a slip by bugging people about blockers more...but it's just a subjective memory, i could be wrong. 15:31:52 cmurf, could add that urgent notification method option and or a cc co-maintainer / un-responsible user option even for things like adamw jsut shouwed with pjones 15:32:26 cmurf: perhaps you could send out a mail to devel@ asking what people think? 15:32:34 ^^ +! 15:32:41 Sure why not? 15:34:51 #action cmurf to mail devel@ for thoughts on blocker notifications 15:35:17 any other f24 notes? should I do an f24 retrospective wiki page? 15:35:52 for non-old-timers - we used to do a thing where we'd put up a wiki page (like https://fedoraproject.org/wiki/Fedora_21_QA_Retrospective ) for thoughts, then turn actionable things into trac tickets 15:36:16 In terms of the end product it's being well received and reliable for the most part so I think that's a good thumbs up for QA. 15:37:22 oh, kparal mentioned the memory relocation error with installs alongside windows 15:37:40 pjones: if you're still around, is there anything we can do to help you there? since we do have a few systems which can reproduce it 15:38:00 If I had any concrete suggestions, I assure you, you'd be hearing them. :( 15:38:07 alright :/ 15:38:25 reverting everything to f23 version is not an option? 15:38:31 until the bugs get resolved? 15:38:41 that would make it worse, not better? 15:38:56 I'm not aware of any critical grub issues with f23 version 15:39:12 pjones: IIRC, reporters of most of the bugs we ran into with UEFI stuff this cycle said the f23 grub2 worked OK 15:39:29 Do we want to do anything special for important bugs that are release blockers? (e.g. bugs affecting composes or that need to land in previous releases for upgrades to work.) 15:39:37 of course, if we try reverting we run a risk of re-breaking things that were fixed at release time... 15:39:58 That suggests most users are UEFI with SB disabled. Because the F23 grub2 chainloader doesn't work when SB is enabled. 15:40:05 because that feature didn't work at all in f23... 15:40:09 right, like that. 15:40:18 pjones: install alongside non-SB windows worked in f23 i think? 15:40:30 Like, I don't get this. The thing that's failing sometimes now did not work at all in f23. 15:40:43 pjones: i thought it did. 15:40:53 the thing that didn't work in f23 was dual boot with SB enabled 15:40:59 which is a big thing, indeed. 15:41:02 Is the relocation failure only with SB disabled? Or other? 15:41:21 cmurf: my case is without SB 15:41:28 not sure about the others 15:41:56 I think what's happened is the deck chairs got rearranged. 15:42:24 F23 GRUB boots Windows if SB is disabled; always fails if SB is enabled. 15:42:45 in my case also SB disabled 15:42:45 Now it's sometimes failing if SB is disabled. 15:42:46 but also there are a pile of other fixes in the f24 package, and I'm reasonably sure reverting several of them would be blockers. 15:42:58 yeah, so i'm not sure it'd be a good idea 15:43:20 #info https://bugzilla.redhat.com/show_bug.cgi?id=1347291 is a worrying issue, but we do not have a resolution for it at present, and reverting to f23 grub2 is not really practical 15:43:32 ok, we're running out of time, so we should probably move on here 15:43:41 #topic Fedora 25 planning 15:44:03 #info current f25 schedule is up at https://fedorapeople.org/groups/schedule/f-25/f-25-quality-tasks.html 15:44:14 adamw, as meta tickets? re retrospective 15:45:12 adamw, re: SB disabled duals did work 15:45:38 i'm not sure how concrete that is, but for now, we have ~1 month to F25 branch point 15:45:57 that's when the release validation treadmill kinda kicks in, though of course we'll have rawhide validation tests before then 15:46:57 Wayland by default attempt again. That's the biggest change I can think of. 15:47:04 sumantro: are you interested in being the official 'in charge of test days' person for this cycle? since you did a good job with them for f24 15:47:17 #info Wayland-by-default is once again on the slate for F25 15:47:21 yes I would love to be! :) 15:47:26 ok, cool 15:47:40 so now would be a good time to send out a 'call for test days' - by email and whatever else, you're good at getting messages out :) 15:47:52 #action sumantro to take charge of test days for f25 cycle, and send out a call for test days soon 15:48:11 Oh, you know what. 15:48:12 sumantro: you can see the f24 test day call at https://lists.fedoraproject.org/archives/list/test%40lists.fedoraproject.org/thread/HC3OGAVTL5Y6OS66AHWI53I2I35AAEVV/ for reference 15:48:14 hm. 15:48:25 i've been basically cut-and-pasting the same mail for several years now :P 15:48:39 Thanks, sure I will kick off the process very soon ! :) 15:48:53 @Adam thanks ! :) 15:48:57 pjones: pins and needles 15:49:22 yeah, added one suggestion to 1347291 that might actually help debug. 15:49:36 on the openqa front we're planning to update openqa to a new upstream snapshot soon, and keep building out more tests before alpha, hopefully we can cover even more of the validation testing this cycle. i'd also like to try and tackle automated testing of updates somehow, at least anaconda updates, but not sure if we'll have time 15:49:44 pjones: thanks 15:50:10 kparal: tflink: do you guys have any big ticket items for f25? where are we with beaker? 15:50:42 automated docker image testing is the big thing that we've committed to 15:51:06 there's some other stuff in the pipeline but I'm not sure it'll be ready by branch 15:51:14 OK 15:51:17 that's in taskotron, presumably 15:51:44 #info openqa plans for F25: new upstream snapshot, extended test coverage, possibly first steps to automated update testing 15:51:53 adamw: yep 15:51:55 #info taskotron plans for F25: automated docker image testing 15:52:24 beaker is in the same place that it has been 15:52:25 does anyone have any other thoughts, ideas, suggestions for where we can go with f25 testing? or big f25 plans that we'll need to work with? 15:52:32 tflink: :/ i see 15:52:38 adamw: good work on networking tests for openQA btw 15:53:01 thanks 15:53:13 adamw: and we should work on reporting too 15:53:22 how do you mean? 15:54:05 I still think that wiki + report email is too little, but that's for another discussion 15:54:28 ah, k 15:54:39 I've been also talking with dgilmore about gating composes by openQA, but that's also for another discussion 15:54:48 alright, we can come back to this in future...let's get to the last agenda item 15:54:53 yeah, i've talked to him about it too 15:55:01 .fas mohamed94 15:55:01 mohamed94: mohamed94 'Mohammed Tayeh' 15:55:06 #topic Onboarding session planning 15:55:08 mohamed94: hi! 15:55:15 hi 15:55:28 hi mohamed94 :) 15:55:35 welcome all 15:55:40 so sumantro has proposed a hangouts (or similar voice+video chat) session for onboarding all the new members he's rounded up recently (welcome new folks, and thanks sumantro) 15:55:49 this seems like a great idea 15:55:56 #info sumantro has proposed a hangouts (or similar voice+video chat) session for onboarding all the new members he's rounded up recently (welcome new folks, and thanks sumantro) 15:56:25 Thanks Adam :) 15:56:29 ah, now i catch up on email, looks like he's sent out the poll 15:56:37 And welcome all! :) 15:56:47 #info poll for timing of QA onboarding meeting is up at http://doodle.com/poll/nbtni4czm8qfwub9 15:56:59 Hey Adam and Sumantro :) 15:57:03 it'd be great for our more experienced members to be around to help with the session and show that we're ready to support new members 15:57:13 so if folks can vote on a time that's convenient for them that'd be awesome 15:57:13 great idea i want to join this session 15:57:15 I sent a poll and an rough agenda with full detailed description! 15:57:25 is anyone really sad about doing it on hangouts? 15:57:39 i'm usually the Official Everything Must Be F/OSS person, but i don't really mind for an informal session like this 15:58:36 hangout is a good but the problem is that closed source 15:58:59 Adam I am too a very deterministic f/oss person , I would love to do it anywhere :) 15:59:05 well then, i guess this is all on track...when i wrote the agenda it seemed like maybe it was a bit stuck but now the poll is up things are moving again 15:59:07 I think it is more important to move forward with something now, rather than delay while we worry about what we'd like to do in the future. 15:59:11 culd try https://meet.jit.si/ ? 15:59:11 brunowolff: yup, for sure 15:59:18 nirik: yeah, someone mentioned that 15:59:20 many folks have had good luck with it. 15:59:26 oh, that's reassuring 15:59:39 i did worry that we wouldn't want the session to turn into a round of 'the tech isn't working for me' 15:59:41 that'd turn people off 15:59:44 I hate hangouts mainly because of what doesn't work, like searches, google voice, rather than the openness. 16:00:08 any of them have problems for some people at some times. ;( 16:00:19 welp, i guess we can pick a time/date and then decide whether to go with hangouts or jitsi or anything else 16:00:30 #action sumantro to gather poll results and schedule the onboarding session at the best time 16:00:31 +1 adam 16:00:40 ok, and we're at 1 hour 16:00:42 so a very quick: 16:00:44 #topic Open floor 16:00:49 any other incredibly urgent business? 16:00:55 Why don't we still use the retrospective page for new releases? 16:01:05 coremodule: it mainly just sort of withered 16:01:22 i.e. it requires someone (i.e. usually me) to put it up and then remember to gather all the responses and turn them into trac tickets 16:01:28 and because i'm an awful person it just kinda didn't happen lately 16:01:35 if you think it's useful, i can put one up for f24 for sure 16:01:45 Oh gotcha. 16:01:46 or if you like, you could take the task :) 16:01:54 Sure, I'd be glad to. 16:02:10 it's not difficult, you just copy/paste the page, create the sub-pages for editing by people without wiki accounts, send a mail to the lists asking people to fill it in, then check back later and make tickets 16:02:13 thanks 16:02:16 I think things run a lot more smoothly than they used to. So there may be less incentive to work on a wiki page to document issues. 16:02:21 #action coremodule to set up the wiki retrospective for F24 16:02:38 I'll post a wiki page in the same format as the F21 one and announce it on #fedora-qa. 16:02:41 Gotcha, I can do that. 16:02:58 brunowolff: true, but of course, the other way of looking at 'running smoothly' is 'stuck in a rut' - so even if we don't need to fix things that are awful, maybe we can make things better if we think outside what we're doing right now :) 16:03:05 coremodule: just ping me if you have any questions 16:03:18 adamw, Sure will. Thanks! 16:03:23 coremodule: you can find the old emails in the archives to use as templates too, you might need to go back to the old mailman2 archive though 16:03:42 adamw, I'll take a look... 16:03:50 alrighty, anything else? 16:04:03 * adamw sets the fuse 16:05:03 thanks for coming everyone! and great job on f24 :) 16:05:15 it's really nice to see so many people pitching in with validation testing and test days and so on 16:05:35 #info thanks to all who worked on f24 validation 16:05:39 oh, yeah... 16:05:46 #action adamw to work on F24 Heroes of Fedora blog posts 16:05:56 * adamw not sure he finished the f23 ones...sigh 16:05:58 #endmeeting