QA/Meetings/20130930

From FedoraProject

Jump to: navigation, search

Contents

Attendees

  • adamw (99)
  • tflink (49)
  • dan408- (40)
  • spstarr_work (34)
  • cmurf (27)
  • kparal (8)
  • zodbot (6)
  • nirik (4)
  • satellit_e (4)
  • kalev (3)
  • roshi (3)
  • brunowolff (2)
  • handsome_pirate (2)
  • spoore (1)
  • jskladan1 (1)
  • mkrizek (1)
  • Cerlyn_ (1)
  • pschindl (1)

Agenda

  • Previous meeting follow-up
  • Fedora 20 Beta pre-TC1 planning
  • Open floor

Previous meeting follow-up

  • tflink to talk to releng about where to file bugs in the cloud .ks and whether it needs to be packaged in the distro - not done yet, tflink forgot. Re-actioned for next week
  • adamw to get F20 common bugs page up and populated - this was done, adam asked for people to add any more appropriate issues
  • adamw to send the update criteria to 'production' - this was done
  • adamw to check if any further criteria changes are needed to beta/final for arm/cloud - adam had taken a quick look, but thought more in-depth assessment was needed

Fedora 20 Beta pre-TC1 planning

  • TC1 scheduled for the following day, 2013-10-01
  • No-one was aware of any major problems which might affect TC schedule
  • Live compose was failing due to a depcheck error, but releng already aware

Open floor

  • tflink noted that there has been talk of having a short schedule for F21
  • spstarr pointed out that KDE may be gaining official Wayland compatibility at that time, which would be a major change
  • We agreed to relay our concerns regarding keeping any changes in a short-cycle release small and manageable to FESCo at their next meeting

Action items

  • tflink or someone competent to talk to releng about where to file bugs in the cloud .ks and whether it needs to be packaged in the distro
  • adamw to talk to anaconda devs and get a build done for beta tc1 with live keyboard bug fixed
  • spstarr to work on iSCSI install testing
  • adamw to look into nightly fail

IRC Log

adamw #startmeeting Fedora QA meeting 15:00
zodbot Meeting started Mon Sep 30 15:00:33 2013 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00
zodbot Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00
adamw #meetingname fedora-qa 15:00
zodbot The meeting name has been set to 'fedora-qa' 15:00
adamw #topic Roll call 15:00
* pschindl is here 15:00
* tflink is present 15:00
adamw #chair tflink 15:00
zodbot Current chairs: adamw tflink 15:00
* Cerlyn_ is here 15:00
* mkrizek here 15:00
* brunowolff is here 15:01
* roshi is here 15:01
* kparal here 15:01
* cmurf is conscious 15:01
cmurf barely 15:01
adamw ahoyhoy folks 15:01
adamw welcome to the uk 15:01
tflink wait, we're supposed to be conscious for the meeting? 15:02
* spoore is lurking 15:02
roshi tflink: it's generally good form 15:02
roshi but not *totally* required I don't think 15:02
spstarr_work hello 15:02
adamw it depends how good your bot is 15:02
adamw mine has been running things here for the last six months 15:02
* adamw arrived in the UK yesterday, still getting settled 15:03
kparal adamw: except for the unexpected breakdown during August/September 15:03
cmurf adamw: Canada kicked you out? 15:03
adamw just on vacation 15:03
adamw well, visiting folks 15:04
cmurf oh ok I thought there was a story. nevermind. 15:04
spstarr_work  :) 15:04
adamw nothing to see here 15:04
adamw continue about your business, citizens 15:04
adamw alrighty, looks like that's everyone 15:04
adamw #topic Previous meeting follow-up 15:05
adamw tflink: "tflink to talk to releng about where to file bugs in the cloud .ks and whether it needs to be packaged in the distro" - anything? 15:05
tflink nope, forgot 15:05
adamw .fire tflink 15:06
zodbot adamw fires tflink 15:06
adamw #action tflink or someone competent to talk to releng about where to file bugs in the cloud .ks and whether it needs to be packaged in the distro 15:06
adamw #info "adamw to get F20 common bugs page up and populated" - this was done: https://fedoraproject.org/wiki/Common_F20_bugs . Do add or mark with 'CommonBugs' keyword any other bugs you think ought to be listed 15:06
spstarr_work lol :D 15:07
adamw #info "adamw to send the update criteria to 'production'" - I went ahead and did this, though can't remember if I sent an email note out about it or not 15:07
adamw so that should be good for beta 15:07
cmurf email went out i saw it 15:07
spstarr_work while you banter I'll grab lunch downstairs quickly :) 15:07
adamw #info "adamw to check if any further criteria changes are needed to beta/final for arm/cloud" - had a quick think about it and couldn't come up with any, but i should check with the arm and cloud folks 15:08
adamw cmurf: a 15:08
adamw ta 15:08
adamw anyone got any notes on those items? 15:08
adamw guess not! 15:09
adamw #topic Fedora 20 Beta pre-TC1 planning 15:09
adamw so, TC1 is due tomorrow 15:09
adamw does anyone know of any crazy emergencies we should be worrying about at this point? 15:09
kparal has somebody actually asked for compose yet? 15:10
* kparal has to find the ticket url 15:10
adamw no, i usually just do it on the day 15:11
tflink just to ask the anaconda devs to build with that live keyboard fix 15:11
adamw no particular reason to file it ahead of time 15:11
adamw tflink: ah, good point 15:11
adamw new anaconda is always a good idea 15:11
tflink people seem to be filing it about once a day 15:11
adamw #action adamw to talk to anaconda devs and get a build done for beta tc1 with live keyboard bug fixed 15:11
adamw yeah 15:11
adamw GNOME 3.10 is done, so that's one thing out of the way 15:13
tflink have we done any raid installs yet? 15:13
tflink or many lvm-thinp installs? 15:13
spstarr_work i dont have enough HD for RAID to test :/ 15:13
adamw i suppose we should focus on GNOME Software testing for beta 15:13
adamw #info GNOME Software testing will be important for beta 15:13
spstarr_work tflink: iSCSI tests? 15:13
tflink spstarr_work: I haven't done any 15:14
cmurf lvmp-thinp is working 15:14
adamw tflink: i didn't :/ 15:14
spstarr_work tflink: I can look at doing some iSCSI testing 15:14
cmurf minus the extra p 15:14
adamw I don't have my RAID test box in the UK so I can't, either 15:14
adamw all i've got here is the sony laptop 15:14
brunowolff You should be able to test software raid with just one HD. It's dumb for normal use, but it should work for testing. 15:14
tflink I have the hw to do raid install tests 15:14
adamw software raid is easy to test in a VM anyway 15:15
adamw it's hw / fw raid that's fiddly 15:15
tflink true 15:15
tflink at this point, I'll probably just wait for beta tc1, though 15:15
kparal tflink: I did some raid tests during Alpha testing, both sw raid and fw raid 15:15
adamw sure, that's not really a showstopper, just something to test at tc1 15:15
kparal http://testdays.qa.fedoraproject.org/testcase_stats/QA_Testcase_Install_to_BIOS_RAID.html 15:16
tflink kparal: did we address infra's hostname concerns for that machine? 15:17
cmurf something that's coming up increasingly is hardware shipping with Intel Smart Response Technology combining SSD/HDD together. Linux cannot support this right now. 15:17
kparal tflink: no idea 15:17
cmurf anaconda doesn't have a test for this, it basically just tells the user they have no disks at all 15:17
cmurf that's in the category of RAID 15:18
spstarr_work #action spstarr to work on iSCSI install testing 15:19
* jskladan1 is late, but present 15:19
adamw cmurf: if we can't support it, then there's nothing much to test... 15:19
adamw cmurf: what could anaconda do instead? can we access the drives 'normally', is it just a 'note that SRT is in use rather than return nothing' case? 15:19
cmurf we can't access the drives normally, but md reports back an imsm attribute that it knows it doesn't support, so anaconda could know this is ISRT and tell the user ISRT isn't yet supported 15:20
cmurf there's a bug around here for md that i bounced back to anaconda for them to look at 15:21
* satellit_e listening 15:22
cmurf . 890881 15:23
adamw spstarr_work: not sure if #action works for non-chairs, i'll do it just in case 15:23
cmurf .890881 15:23
adamw #action spstarr to work on iSCSI install testing 15:23
adamw cmurf: worth keeping an eye on, then, but not world-ending 15:23
cmurf aparently the bot doesn't know what period means today 15:23
cmurf https://bugzilla.redhat.com/show_bug.cgi?id=890881 15:23
tflink isn't there a ssd cache test day scheduled? is that different than the isrt? 15:24
cmurf no it's just an fyi about raid. when a vendor like HP ships a pile of hardware with it on by default… we're going to see users wondering why the installer shows them no drives and no error messages. 15:24
cmurf tflink: it's the same general idea, as isrt, different implementation 15:25
adamw tflink: I think that's the Linux in-kernel version of SSD caching 15:26
cmurf there are two, dm-cache and bcache 15:26
cmurf i don't see anaconda options for them though 15:27
adamw of COURSE there are. 15:27
cmurf maybe ping Rolf and find out what the anaconda plans are or are not for F20? 15:27
cmurf https://fedoraproject.org/wiki/Changes/SSD_cache 15:27
adamw so, any other Beta business? it's looking more or less like 'steady as she goes' 15:28
adamw i guess i'll do a live compose test today, or just grab the latest nightly and try that 15:28
spstarr_work thanks 15:28
tflink adamw: nightlies failed last night 15:28
adamw ah, worth looking into then 15:29
tflink releng is aware of the issue, dep conflict 15:29
adamw #action adamw to look into nightly fail 15:29
adamw ok ok 15:30
nirik should be fixed tomorrow 15:30
adamw .fire nirik 15:30
zodbot adamw fires nirik 15:30
adamw TOO LATE 15:30
spstarr_work tomorrow isn't today 15:30
spstarr_work you're too slow :P 15:30
adamw darn right 15:31
adamw okay, moving on! 15:31
adamw #topic Open floor 15:31
spstarr_work continuous integration requires rapid development and testing :P 15:31
tflink as a heads up, there is noise of picking up the pace for f21 15:32
cmurf bcache support for anaconda for F21. Just saw the email. 15:32
tflink ie, having a "small" release and getting back on track for a may release 15:33
spstarr_work 'small' ? 15:33
* satellit_e FYI soas build Fedora-20-Nightly-x86_64-Live-soas-20130925.08-1.iso fixes https://bugzilla.redhat.com/show_bug.cgi?id=1008569 15:33
tflink yeah, what f20 was supposed to be 15:33
spstarr_work considering KDE i hope will be closer to Wayland by 21... that wont be small 15:33
adamw so, again with that plan 15:33
adamw i thought a few weeks ago people were talking about doing a *longer* cycle 15:33
tflink yeah, but the fedora.next proposal is taking time to come together 15:34
adamw wayland is probably a good thing to keep in mind in making that decision, yeah... 15:34
tflink the thought process is that there's no point in delaying a release until the various WGs have formed and figured some stuff out 15:34
spstarr_work adamw: indeed 15:34
tflink as always, there is pressure from gnome to get back to the old cadence 15:34
* handsome_pirate comes in just in time for open floor 15:35
tflink tickets should be filed with fesco soon 15:35
adamw #info tflink notes there's talk about a short F21 release cycle (again) 15:35
* handsome_pirate apologizes for getting distracted by work 15:35
spstarr_work tflink: but gnome can't dictate Fedora alone 15:36
spstarr_work if we treat KDE and MATE and others fairly 15:36
tflink it sounds like there will be a call for proposals on what we would do with a delayed release 15:36
tflink spstarr_work: I'm simply reporting what I've been hearing. not here to argue about which de does what 15:37
tflink reporting/repeating 15:37
spstarr_work tflink: ya 15:37
adamw yeah, this is more impact on QA than arguing about the idea 15:37
adamw if there is a plan for a 'small', short release, what do you folks think of proposing we go with a 'two release point' cycle? 15:38
adamw beta and final only 15:38
adamw i'd be worried about trying to cram alpha, beta and final into 4 months or less 15:38
tflink if we wanted a delayed f21, we'd have to counter the argument that there's little point in investing in infra/qa before we know where the 1 -> 3 product thing is going 15:38
* kalev thinks it's an excellent plan. 15:38
tflink adamw: but it's a small thing, that's no problem 15:38
tflink kalev: I think most of the people pushing for it are not as heavily affected by the short turnaround 15:39
tflink for the record, my comment about being a small thing should have had <sarcasm> tags on it :) 15:40
kalev tflink: no, I mean it's an excellent plan to drop alpha to make it easier to shorten the cycle 15:40
kalev easier on QA mostly. 15:40
tflink kalev: if you're not the one testing this stuff, sure 15:40
tflink dropping alpha on a release that would likely have more wayland stuff sounds like a bad idea to me 15:40
spstarr_work adamw: but short release means no changes to Anaconda..? 15:40
adamw well. that depends on what the definition of 'small' winds up being. 15:41
adamw if this plan has legs, we'd definitely want to push for 'small' to be serious. i.e. no mass migrations to wayland, no major anaconda changes. 15:41
tflink adamw: the question that comes to my mind is how to get that. commitments from devs? 15:43
cmurf there is X amount of work for a release whether it's small or large so i don't see much point in a small release 15:44
cmurf a lot of work for not a lot of gain 15:44
dan408- hi 15:44
adamw tflink: fesco laying down the rules at time of approving schedule? 15:44
dan408- finally made it to a meeting 15:44
adamw usually if fesco out and out says "This Shalt Be So", that's what happens 15:44
adamw cmurf: the point is to get us back on our usual schedule, specifically its tie-in with GNOME and with various holidays 15:45
dan408- is the plan to shorten f20 or f21? 15:45
tflink but if we're delaying/dropping a release, I'm not sure I see much utility in stepping up the pace for another relase 15:45
tflink dan408-: 21 or 22 15:45
tflink dan408-: sorry, wrote too quickly. shortening 21 15:46
dan408- might as well announce the shortening of 21 asap before it's too late 15:46
tflink I suspect that it'll be a topic @ the fesco meeting on wednesday 15:46
dan408- right 15:46
spstarr_work fesco should wait until wayland is integrated in the ecosystems first 15:47
tflink didn't stop us from getting caught a bit flatfooted for f20 15:47
dan408- if we limit the number of "changes" it shouldnt be a big deal 15:47
tflink s/us/me? 15:47
dan408- is gnome 3.10 complete? 15:47
tflink it was released 15:47
spstarr_work i dont want there to be egg on face if Fedora 21 is cut before KDE has wayland support (assuming its close to that timeframe) 15:47
dan408- that doesnt mean it's complete :P 15:47
tflink I assume so :) 15:47
dan408- so they are alraedy starting on 3.12 then? 15:47
dan408- spstarr_work: is KDE also on a 6 month release cycle? 15:48
spstarr_work dan408-: well, thought so, but wayland support might take longer(?) 15:48
dan408- well 15:48
spstarr_work KDE '5' is coming and i'd hate Fedora 21 to just miss it 15:49
adamw spstarr_work: well, i'd sure count a KDE migration to wayland as a 'major change' 15:49
dan408- i think talking about this here is pointless, because fesco is just going to decide what they're going to decide anyway 15:49
adamw i'd feel pretty uncomfortable trying to do that in a four month cycle. but eh 15:49
dan408- moreover, I also think basing our entire release schedule based on 1 DE's schedule is stupid. whether it's gnome or KDE, that's just my opinion 15:49
tflink it sounds like it'd be wise to keep an eye on fesco tickets 15:49
adamw dan408: the point is to determine what our concerns about the possibility are and relay those to fesco, and to think ahead about what we need to do ourselves to deal with such a schedule 15:50
tflink and plan for a break in the blocker review meeting wednesday 15:50
dan408- we'll get delayed again some day 15:50
spstarr_work adamw: they have early support in KDE 4.11 for wayland but its not even tech preview level 15:50
dan408- adamw: from a QA perspective, will we have enough time to do all the proper testing? I think yes. 15:50
dan408- okay awkward silence.. 15:52
adamw tflink: point. 15:52
spstarr_work adamw: I want to think Fedora should move to a new FESCO model 15:53
cmurf so beta tc1 tomorrow? 15:53
adamw #agreed we have concerns about a short release cycle especially combined with potential upcoming features like official wayland support for KDE and/or GNOME, we will raise those with fesco 15:53
adamw spstarr_work: out of scope. 15:53
spstarr_work representation from QA (1 person) and one from infrastructure etc... 15:53
spstarr_work vs some people who are not privy to QA even 15:53
adamw #info adamw suggested the possibility of doing a beta/final only schedule for a short release cycle, tflink was worried about dropping alpha 15:53
adamw spstarr_work: still out of scope. 15:53
spstarr_work adamw: we are open floor though :) 15:54
dan408- adamw: when did the meeting go to 8am? 15:54
tflink I'm not against dropping alpha if we are actually doing a small release 15:54
cmurf floor isn't that big 15:54
adamw dan408: it hasn't changed... 15:54
dan408- hmm 15:54
adamw spstarr_work: it's open to all QA topics, not general fedora proposals :P 15:54
adamw tflink: fair enough 15:54
dan408- oh 15:55
dan408- one thing i'd like to raise that's QA related is regarding the spins 15:55
adamw dan408: did the clocks change already in your region or something? 15:55
dan408- no 15:55
dan408- im just never awake at 8AM or I'm getting ready for work 15:55
adamw what's up with the spins? 15:55
spstarr_work adamw: fair enough 15:55
dan408- ok regarding spins, a bunch of spins almost werent shipped for alpha 15:55
dan408- there is a new signoff process for spins 15:55
adamw yeah, it's been brought up before i think 15:55
dan408- so it is paramount that this gets done before the alpha/beta/final composes 15:56
tflink sure, by the sigs responsible for the spins 15:56
dan408- LXDE, Scientific, MATE, and another one almost didnt get shipped 15:56
dan408- tflink 15:56
dan408- no QA is responsible for this 15:56
tflink no, we're not 15:56
dan408- we're not responsible for testing? 15:56
dan408- alright then 15:56
tflink the sigs are 15:57
tflink for the signoff matrices 15:57
dan408- i dont think that's been properly communicated out 15:57
satellit_e +1 15:57
tflink iirc, it was part of the matrix announcement 15:57
dan408- whatever, i know now for my spin 15:58
kparal dan408: just read once again nirik's announcement. it has been posted into every major list 15:58
* nirik is open to ideas for more communcation. 15:58
satellit_e link should be on all testing pages https://fedoraproject.org/wiki/Releases/20/Spins 15:59
dan408- hmm 15:59
nirik if you can give me home addresses of spin owners I can send them all a flaming engraved candygram? ;) 15:59
adamw dan408: yeah, the idea is that if you want your spin shipped, you get it tested 15:59
dan408- adamw: it's not just me. 15:59
adamw spin owners could come and ask us nicely to help test, of course, but the process doesn't require us to or give us ownership of it 15:59
adamw dan408: 'you' in the general sense. 15:59
adamw i.e. it's the spin owner's responsibility. 15:59
dan408- nirik: pm 15:59
dan408- nirik: lol 15:59
adamw okay, looks like we're good 15:59
spstarr_work adamw: I'd hope the spin owners know, I'd hate to know if KDE somehow got missed in a release, but I suspect they do very well know 15:59
* adamw sets quantum fuse 15:59
nirik  :) 16:00
tflink spstarr_work: kde and desktop are the exceptions 16:00
adamw we are responsible for KDE testing as it's a elease blocking desktop 16:00
spstarr_work tflink: oh? 16:00
adamw KDE and GNOME we're on the hook for 16:00
spstarr_work  :) 16:00
dan408- i'll probably be propsoing 2 more spins for f21 16:00
* spstarr_work smile 16:00
spstarr_work I do KDE side easier as I use KDE 16:00
dan408- adamw: double standard ;p 16:01
dan408- anyways we luckily dodged a bullet on that one 16:01
adamw alrighty, thanks for coming folks 16:02
dan408- thanks for hosting 16:03
adamw yoooooooou're welcome 16:03
adamw #endmeeting 16:03

Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!