16:00:33 #startmeeting Fedora QA meeting 16:00:33 Meeting started Mon Mar 2 16:00:33 2015 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:33 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:36 #meetingname fedora-qa 16:00:36 The meeting name has been set to 'fedora-qa' 16:00:41 #topic Roll call 16:00:48 * kparal is here 16:00:52 ahoyhoy folks, who's around for some qa...ing 16:01:01 * nirik is lurking if he can help with anything 16:01:04 * Kinokoio is here 16:01:09 * satellit listening but have to leave early 16:01:14 * jreznik is here 16:01:22 * tflink is here 16:01:31 * pschindl is here 16:02:49 * smccann is here 16:02:56 * pwhalen is here 16:05:38 good turnout :) 16:05:55 #chair pschindl satellit 16:05:55 Current chairs: adamw pschindl satellit 16:06:08 how's everyone doing so far this cycle? having fun? 16:06:38 always :) 16:07:20 having fun is the only way how to survive sad world :) 16:07:56 hehe 16:08:06 wow, we're off to a cheerful start :) 16:08:06 * satellit_e does coconut report warnings or failed on Test results page? 16:08:12 'hey, if i squint a little bit these tears look like laughter!' 16:08:18 satellit: not yet, it only reports passes 16:08:23 k 16:08:28 satellit: so far we figure it's not reliable enough to report fails 16:08:48 (sometimes its failures are simply bad test cases or bugs in autoqa or small changes in the installer's appearance or whatever) 16:08:56 er, openqa :) 16:09:26 i'd like to teach it to recognize the crash screen and file a bug when it sees it, that might be on my project list for this week 16:09:41 nice 16:09:57 ok, let's have a quick impromptu coconut topic since it came up 16:10:01 #topic Project Coconut / OpenQA 16:10:09 * Kinokoio is going to reboot, will leave kinokoio_mobile for a while 16:10:21 * roshi is here 16:10:32 #info Coconut (the openQA-based automatic testing thing) does not report failures at present as they're too often false alarms 16:10:56 satellit: i have my own instance of coconut up at https://openqa.happyassassin.net/ now 16:11:34 satellit: i'm planning to have it run the test suite regularly (haven't set that up yet but I will), so you'll be able to look through its recent runs and maybe spot interesting failures and reproduce them for bug filing 16:11:53 ok 16:12:25 #info adamw has his own experimental instance of Coconut up at https://openqa.happyassassin.net/ where people can see the system in operation and maybe look for test fails that can be reported as bugs 16:12:38 a lot of the time failures are just going to be me messing around, though :P 16:13:08 anything else on the topic while we're at it? 16:14:07 kparal: jskladan: any interesting news on your end? 16:14:23 nope, not really 16:14:30 okey dokey 16:14:32 * kparal is only ankle's deep in openqa testing 16:14:39 i'll send a PR for my firewall test(s) at some point 16:15:08 can anyone submit tests to your instance adamw ? 16:15:31 roshi: no-one but me at present, sorry 16:15:57 i haven't entirely looked into the mechanics of safely granting perms to others yet 16:15:57 np, just wondering if it was a means to get ones feet wet with it 16:16:13 roshi: you can't really do much interesting messing around from the web UI only 16:16:20 it's mostly meant as a result-viewing tool i think 16:16:36 I didn't figure - just wondered if there was some way to submit testcases 16:16:37 hi everyone 16:16:40 hi amita 16:16:40 which are in perl, right? 16:16:43 o/ 16:16:55 roshi: so what you can do is check the fedora ones out of git: 16:17:13 https://bitbucket.org/rajcze/openqa_fedora 16:17:27 and poke around from there...then ask me or jskladan to test them i guess 16:17:34 or at least add them into our instance for testing 16:17:38 it's all a bit ad hoc atm :) 16:17:56 fair enough 16:18:06 roshi: the tests are technically perl but really it's mostly a restricted subset which isn't horribly insane 16:18:21 i had to go and learn some perl to submit a bugfix for openqa itself though. i'm still getting over the trauma. 16:18:25 as long as one sticks with the basics 16:18:58 * roshi hands adamw a cold one 16:19:06 * satellit_e it might be nice to mark which test coconut has tried and not passed I see it progresses thru them 16:19:08 the hell is with all this 'shift' and '@_' crap :8 16:19:15 we'll get through this adamw :p 16:19:18 alllllrighty 16:19:20 :P 16:19:27 let's not take up too much time, moving along! 16:19:45 satellit_e: the thing is, that "not passed" is "failed test result" as often as "failed test execution" 16:19:51 so I'm not really willing to go there 16:20:02 #topic Previous meeting follow-up 16:20:09 We can mark which testcases coconut tries to test, though, in some way 16:20:12 but not with results 16:20:17 +1 16:20:46 jskladan: we could actually do something fairly easily with extra {{result states 16:20:49 i was already thinking about that 16:20:50 * satellit_e like koji 16:21:16 one thing i was thinking of was having the template know which users are bots and indicate that 16:21:26 so results from coconut and colada would have 'BOT' in bold stuck in them somewhere 16:21:51 and we can easily add a state like condfail or possfail or something, just a case of coming up with a sensible way to render it visually 16:22:49 the nice thing is all the tools run off the wikitext, so we have leeway to fiddle around with the appearance of the rendered template without screwing up wikitcms or anything 16:23:37 aaanyhoo. i really ought to've planned this meeting topic before we got here, sorry folks - let me check the last meeting log... 16:24:33 hum, looks like we didn't really have much to follow-up on from last time 16:24:48 am i missing anything? 16:25:00 * danofsatx has arrived 16:25:16 hey danofsatx 16:25:31 can't think of anything from last meeting adamw 16:27:16 okely dokely 16:27:18 morning dan 16:27:35 #topic Fedora 22 status 16:27:53 so we have blocker review coming up in a half hour and we'll go into things there too, but always worth doing a check-in... 16:28:04 it feels like we spent the whole of last week firefighting showstoppers, right? :/ 16:28:13 for sure 16:28:33 seeing i386 lives problems in boxes not sure it is real need more testing 16:28:58 we got networking working in cloud images again 16:29:00 * satellit_e install and sometimes boot 16:29:03 how's latest TC is looking? 16:29:16 satellit: that's the known bug IIRC 16:29:21 TC7 doesn't have every last fix we figured out 16:29:28 I'm planning to fire TC8 / RC1 today 16:29:34 but direct kernel boot doesn't work anymore in the cloud images 16:30:04 roshi: oh fun :/ 16:30:12 satellit: you're very likely hitting https://bugzilla.redhat.com/show_bug.cgi?id=1197224 , which isn't fixed in TC7 16:30:26 thanks 16:30:32 i figured people must be getting TC fatigue and it'd be kinda rude to spin a TC8 with just that fix :P 16:30:38 so I have a quick topic... what are the chances of us getting a xfce update in before alpha? ;) I guess if rc1 goes today not too good. 16:31:12 well, personally i'm not hugely opposed because hey, if xfce breaks it's not like we slip...but if you file/propose an FE we can discuss it in blocker review? 16:31:26 yep. we hope to do builds today and create the update... 16:31:42 and we need comps changes so it wouldn't happen until tomorrow. 16:31:47 at best 16:31:51 ok 16:31:59 but we will do what we can 16:32:38 roshi: so how far are we from full coverage for at least Alpha testing for cloud? 16:33:22 well, since networking works - pretty close 16:33:29 we haven't covered all the server testing yet which worries me a bit 16:33:33 the only issue is local testing at this point 16:33:43 if anyone's able to look at the non-covered server tests that'd be a big help 16:33:53 since most people use testCloud or the docs it's based on and create a seed.img to provide the metadata 16:33:54 the tc7 image didn't come up right in our old cloud or our new one 16:34:14 roger, I'll take the Server tests 16:34:21 at least the ones I can do, anyhow 16:34:22 thanks dan 16:34:25 yeah, we still need to look into that - I have reports that it does work in some flavors of openshift 16:34:30 i'm gonna try and get some done too, just the more the merrier... 16:34:50 something isn't letting dracut see /dev/vda1 during boot for direct kernel boots 16:35:00 we are coming up on go/no-go, so we definitely need to be aiming to cover any tests we haven't done yet 16:35:03 and I suspect we'll need some discussion as to if that's blocking or not 16:35:10 is it on the proposed list? 16:35:15 thanks for go/no-go reminder adamw 16:35:25 no, I need to file a bug for it 16:35:31 yeah, get it in there :) 16:36:18 will do 16:36:24 going to be a longer meeting today 16:36:40 yeah, looks that way :/ lots of fun bugs 16:36:54 anyone have any other f22 discussion topics? 16:37:09 oh, and remember about testcase_stats , folks - https://www.happyassassin.net/testcase_stats/22/ 16:37:24 it was broken for a little while but it's working now - use it to help spot the tests we didn't run yet (or for a long time)( 16:37:50 * danofsatx was just going to ask for that link 16:38:28 just the yum/dnf migration 16:40:37 yah, that has its own topic 16:40:40 which we're coming to...now! 16:41:01 #topic Yum -> DNF test planning 16:41:09 so many thanks to roshi for doing a preliminary overview of the scope of testing required for this 16:41:45 #info roshi's overview of the scope of the yum->dnf migration for testing purposes can be seen at https://lists.fedoraproject.org/pipermail/test/2015-February/125166.html 16:41:50 apologies for not replying to it yet :) 16:42:34 no worries :) 16:42:49 so we have to decide what our next move is 16:42:59 jskladan: nirik: where is fesco at with the Change process right now? 16:43:11 has the 'broadly complete and testable' check been done yet? 16:43:27 for dnf? not sure, but I thought it was supposed to be all ready 16:43:29 adamw: I do not really know, maybe jreznik does ;) 16:43:36 adamw: https://fedorahosted.org/fesco/ticket/1416 16:43:48 the only outstanding question was the dnf-yum thing, and they aren't going to do that part. 16:43:49 it's after review 16:43:55 the change likely needs updating to note that 16:44:38 jskladan: me? why me? :D 16:45:32 so as things stand the Change is going ahead as planned, right? 16:45:46 AFAIK, except with no dnf-yum 16:45:51 (by default) 16:45:54 ok. 16:46:15 and I assume dnf python2... not dnf-3... 16:46:17 so i guess the first question is whether we're worried about that given the scope roshi identified 16:46:53 but rawhide/f23 has moved to dnf-3 16:47:00 or whether we want to focus on implementing as much of the test plan as possible 16:47:13 roshi: what's your feeling? 16:49:04 well 16:49:11 it's a lot to test 16:49:46 my guess would be to split up areas and have people focus down on those areas 16:49:52 divide and conquer, if you will 16:50:07 sounds reasonable 16:50:19 assuming that the scope I laid out is, you know, correct :p 16:50:27 anaconda testing looks like we're mostly going to hit in the course of regular validaiton testing 16:50:34 I'm assuming that I missed something that's going to come back and bite us 16:50:40 (and anything in there that *isn't in the validation tests probably should be) 16:50:41 but that's a standing assumption 16:50:46 roshi: oh, that's always entirely safe to assume, yeah :P 16:51:16 to me abrt looks like something we could hit nicely with a test day 16:51:32 for our case, are we sticking with my interpretation of which bits QA is responsible for? 16:51:40 (release images and updates) 16:52:03 i think it's reasonable to say those are the bits we're going to primarily *concerned* with 16:52:30 if you get down to theoretical brass tacks about what we're allegedly 'responsible' for it might be more than that, but it's kind of pointless if we're clearly not going to have the reosurces 16:52:45 so i'd say we try to cover this stuff first, then we can worry about more stuff if we have time, which we inevitably won't. :P 16:52:48 * satellit_e afk 16:53:00 does that sound reasonable? 16:54:21 sounds reasonable to me 16:54:52 reasonable aye 16:56:03 * kinokoio wonders where to start 16:56:25 which ever bit is closest to you kinokoio :p 16:57:17 anyone have any other thoughts on this? 16:57:43 roshi: maybe a good next step would be to break it down into actionable things, as trac tickets or phab or anything else you're comfortable with, and we can work on doing them from there 16:58:28 sounds good 16:58:50 #action roshi to break down the migration testing into doable chunks 16:59:00 i don't think you have a chair do you? 16:59:01 actually, I'm not chair I don't think 16:59:05 #action roshi to break down the migration testing into doable chunks 16:59:15 * adamw still following the time-honored jlaska Random Chair Throwing tradition 16:59:36 also, before we quit out - kparal had timezone questions about the slew of meetings we have monday mornings 16:59:53 right 16:59:54 * kparal was waiting for open floor 16:59:56 let's get to open floor quickly 16:59:59 #topic Open Floor 17:00:09 +1 have blocker meetings follow DST changes 17:00:16 here's the request: https://lists.fedoraproject.org/pipermail/test/2015-March/125277.html 17:00:20 since they're lined up with the other meetings it just makes sense. 17:00:38 +1 17:00:46 I also prefer following local time 17:01:08 we'll just have to make sure we switch all three at the same time 17:01:18 +1 17:01:18 (since the actual dst switchover points tend to vary between countries) 17:01:54 anyone opposed? 17:01:56 true, but I also think we should just pick something and stick with it 17:02:06 tflink: how do you mean? 17:02:28 adamw: one time to switch, which I suspect is what you meant 17:02:29 we always followed north america summer switch time for QA meetings, so I think it makes sense to stick with it for all meetings 17:02:45 ah i see. yes 17:03:12 makes sense to me 17:03:16 I think we've usually followed NA DST in the past (I think all of NA switches at the same time) 17:03:16 even within NA I'm not sure it's the same everywhere...but let's just say whoever announces a meeting with the switch first, we switch all the others at the same time :) 17:03:22 * roshi hates the spring timechange 17:03:39 and try to get that week's announcements out nice and early 17:03:44 (like i always don't) 17:03:54 all our meetings except blocker bug are tied to New York timezone in fedocal 17:04:01 I can adjust blocker bug meeting to the same 17:04:15 sounds fine 17:04:25 just make sure to mention it in the next announcement 17:04:26 thanks 17:04:32 #agreed blocker review meetings will move with DST changes as the other two meetings do 17:04:52 #action kparal to adjust fedocal to tie blocker review meetings to New York DST (as the other two are) 17:05:02 alrighty, anything else? if not let's move right on over to #fedora-blocker-review :) 17:05:15 * adamw sets the Acme Necessarily Short Fuse 17:05:24 I propose a 10 minute coffee break before we start the blocker review 17:05:32 since I have to make coffee :) 17:05:32 pfeh, if you must :P 17:05:37 so XX:15 ? 17:05:53 ack 17:05:58 i'm game for :15 17:06:09 and we can give warning to other channels since I haven't yet 17:06:13 wfm 17:06:25 blocker review is in this channel? 17:06:39 kinokoio: #fedora-blocker-review 17:06:45 yeah ^^ 17:06:52 roshi types faster than me 17:07:12 alriiiighty then 17:07:13 thanks 17:07:52 * adamw wonders if it's possible to draw a point in time at which the cultural zeitgeist switched from Carrey to McConnaghoweveryouspellit 17:08:24 * danofsatx looks askew at adamw 17:08:35 danofsatx: 'well alllriiiiighty then!' to 'alright alright alright' 17:08:51 IT IS THE KEY QUESTION OF OUR TIMES 17:09:10 if you say so. Is the fuse set yet? 17:09:53 * adamw checks fus- 17:09:54 * kparal has adjusted fedocal 17:09:56 #endmeeting