QA/Meetings/20120604

From FedoraProject

< QA | Meetings
Revision as of 01:08, 5 June 2012 by Adamwill (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Contents

Attendees

Agenda

  • Previous meeting follow-up
  • Fedora 18 planning
  • AutoQA update
  • Open floor

Previous meeting follow-up

  • N/A

Fedora 18 planning

  • Anaconda new UI:
    • clumens plans to merge the new anaconda UI into Rawhide as soon as custom partitioning is implemented and broadly works. He estimates one month for that
    • We don't believe QA can achieve much practical testing until that merge point, but we asked clumens to provide one or two public test images before then so testers can familiarize themselves with the new UI
    • Key areas of installer that provide much of the testable complexity are repository configuration, storage configuration and bootloader configuration
    • If newUI lands in anaconda significantly before F18 branch point, we will ask releng to do rawhide nightlies until branching
  • Fedora 17 retrospective tasks - Adam will work on this
  • Housekeeping tasks for 16, 17 and 18 cycles - Adam will sync with rbergeron

AutoQA update

  • No AutoQA news, team still in F17 recovery

Action items

  • adamw to review F17 retrospective and come up with an action plan
  • adamw to work with rbergeron to make sure housekeeping gets done for f16/f17/f18

IRC Log

adamw #startmeeting fedora-qa 15:01
zodbot Meeting started Mon Jun 4 15:01:05 2012 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01
zodbot Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01
adamw #meetingname fedora-qa 15:01
zodbot The meeting name has been set to 'fedora-qa' 15:01
adamw #topic roll call 15:01
* red_alert :) 15:01
adamw morning folks 15:01
* satellit_ listening 15:01
* adamw goes to get breakfast 15:01
* tflink is here 15:01
* mkrizek is present 15:02
* akshayvyas is here 15:02
* pschindl is here 15:02
* aalam is here 15:02
adamw alrighty 15:04
adamw we may have a clumens in a minute 15:04
* maxamillion is here 15:04
clumens already here. 15:04
adamw aha 15:05
adamw you lurker 15:05
adamw awesome 15:05
adamw so we can actually skip previous meeting follow-up as there's nothing from the last meeting 15:05
* kparal is here 15:05
adamw and zap straight on to... 15:05
adamw #topic Fedora 18 planning 15:05
adamw sooo...yeah. everyone stop relaxing about F17 and start panicking about F18. 15:06
maxamillion lol 15:06
maxamillion I'm excited about F18 15:06
adamw we'll have more planning as we go along of course, but i wanted to get one session in early to try and figure out where we're at with the biggest feature we'll have to deal with - the anaconda UI re-design 15:07
adamw for anyone who's not yet aware: anaconda in f18 will have a completely new UI. see https://fedoraproject.org/wiki/Features/NewInstallerUI for details. obviously, this will be heavy on QA work. 15:07
* brunowolff is here for a little bit 15:07
adamw equally obviously, we're gonna want to start testing it ASAP 15:07
brunowolff Good luck. There is a problem with build initramfs right now. 15:08
adamw so...clumens, what's the current status? i think i recall recently seeing someone say they'd just done the first successful install to HD with newui, so it's obviously still somewhat in the works :P 15:08
clumens well, you can install via at least one method to previously blank disks using autopart 15:09
clumens and several of your customizations will likely even be reflected in the installed system 15:09
adamw heh 15:09
* j_dulaney waves 15:10
adamw hi dulaney 15:10
clumens also, dlehman was talking about having done the first successful hdiso install. 15:10
adamw ah. 15:10
red_alert being able to install is nice, but the more important question is probably: at what point does it make sense to track down and report bugs? :) 15:11
adamw but i guess we're still not to the point where outside testing is going to tell you a lot you don't know? you're still at a point where only a few targeted paths work? 15:11
clumens there's this, which needs to be updated: http://fedoraproject.org/wiki/Anaconda/TestableUI 15:11
adamw red_alert: get outta my head 15:11
adamw right, i figured it needed an update 15:12
red_alert adamw: but it's comfy! so much space to stretch my legs ;D 15:12
clumens i am targeting merging for when custom partitioning basically works, regardless of the state of things like your timezone being right. 15:12
clumens and i'm working on custom partitioning now 15:12
adamw okay. yeah, i was gonna say, at a broad stroke, most of the testable complexity comes with a) repositories, b) storage, and c) bootloader configuration 15:13
j_dulaney Does Mo have a UI for that? 15:13
adamw so it'd be good to know the timeframe on those bits 15:13
clumens repo stuff is pretty far along. 15:13
clumens there's no ui for adding additional repos, but i don't consider that an impediment. 15:13
clumens storage is gonna take more time, but it doesn't have to be feature complete before i want people looking at it. 15:14
clumens and bootloader... good question. 15:14
adamw red_alert: why, i oughta... 15:14
tflink are there any publicly available images for testing? 15:15
clumens if you're on the RH network, yes 15:15
adamw so, no 15:15
adamw  :P 15:15
clumens by which i do not mean RHN 15:15
akshayvyas tflink i dnt think so 15:15
maxamillion clumens: dare I ask why its in RHN? 15:16
adamw maxam: "by which I do not mean RHN" 15:16
maxamillion oh 15:16
maxamillion heh 15:16
* maxamillion fails 15:16
maxamillion I read that as "by which I do mean RHN" 15:16
adamw i'm not hugely concerned about test images until testing can actually contribute something 15:16
clumens i will never mean RHN. 15:16
adamw clumens: on bootloader, has that been whiteboarded out and so on, and just waiting to get implemented? 15:17
clumens yeah i just don't want that point to be too late. 15:17
maxamillion clumens: yeah, I wouldn't think so :) 15:17
adamw clumens: except in contexts such as 'i call fire down upon...' 15:17
maxamillion adamw: +1 15:17
clumens adamw: there's nothing for bootloader yet. 15:17
adamw ookay. that gets me a little worried, since we know what a messy little area it can be. 15:17
clumens well, how much testing does the current bootloader UI get? 15:17
clumens with all the adding args and passwords and crap? 15:18
j_dulaney A fair bit 15:18
clumens we can probably think up something pretty easily to just specify which device gets the bootloader. 15:18
adamw clumens: it's not so much that added credit stuff i worry about as just the bush of messy problems with bootloader _location_ that we unearthed through f15 and f16 with the grub2 switch 15:18
clumens a stop-gap, if you will. 15:18
adamw clumens: that's the most important bit, yeah. 15:20
adamw we already have a hacky stopgap in f17, i was hoping it would get better with newui, but ah well 15:20
adamw once it's in there, we can test it and figure out where it causes issues 15:20
clumens my concern right now is getting something in for F18 that's good enough. 15:21
clumens if bootloader doesn't suck more than it did in F17, i can live with that. 15:21
adamw okay. we can talk about the design outside of qa meeting, as long as we know it's a key thing to have in place for testing that's enough 15:22
adamw can you guess a rough timeframe on when you'll hit the merge point (custom partitioning in place)? 15:22
clumens a month? 15:22
adamw so if we double that on general principles, it's right around alpha candidate time, which isn't terrible 15:23
clumens that would certainly be nice if it worked that way. 15:24
adamw it might be nice to have one or two publicly available images ahead of that time, just so non-RH testers can get a look at what's coming and maybe spot any huge missing areas 15:24
clumens i know of a couple already (like, disks without labels get completely ignored) 15:24
clumens but yeah. 15:24
clumens maybe i'll try to come up with a solution to that today too. 15:25
clumens i could try to have something publically available mid-month. 15:26
clumens i'm out all next week, so... 15:26
adamw sorry, got lagged out, i just caught up from logs. mid-month sounds great. 15:26
adamw for qa folks, we'll be wanting to try and get through the whole anaconda test suite as soon as we can, so we'll probably be looking at doing more testing at alpha stage than we usually do 15:27
clumens i need to deputize someone to deal with patch review and image building in my absence. 15:27
clumens how much of the test suite has language that's going to need to be changed for this? 15:27
adamw that's a good question, i don't know the answer offhand, anyone? 15:28
clumens also: you would probably want a publically available image to not have loader in it, i assume. 15:28
adamw the intent of tests is usually clear even where they document the current UI, so it shouldn't be a huge hurdle. 15:29
adamw for the purposes of looking over the new UI, loader presence/absence doesn't matter much 15:29
adamw once we get to the point of active testing for specific bugs, yeah, we'll want to be using noloader images. 15:29
clumens good, because i really don't want to spend my time dealing with git merge. 15:29
adamw we have tests like hardware RAID which read "At type of devices step, choose Specialized Storage Devices and click Next.", "Select hardware RAID device and click next." etc. of course that'll need updating. 15:30
clumens yes. 15:30
adamw at a wag, just looking through the list of test cases, i'd say maybe, oh, a half of them will need tweaking. but it's the kind of thing we'll get done in a couple of days once we know what the UI looks like. 15:31
clumens the UI looks almost exactly like the mockups 15:31
adamw well, that's good :) 15:32
adamw okay. i think we're all pretty much on the same page now, right? does anyone have any other questions about newUI as regards qa planning? 15:32
red_alert could we maybe set up nightly builds for rawhide as long as F18 branching hasn't happened yet? 15:33
red_alert instead of having clumens create new images manually every other day :) 15:33
tflink red_alert: wouldn't it be wise to wait until things are testable? 15:34
tflink unless I'm missing something here (which is very possible) 15:34
clumens i guess you would have to ask rel-eng for that sort of thing. right now newui code isn't even on the master branch of anaconda, so it's nowhere where real builds get pulled from. 15:34
adamw once anaconda's merged, then there already _are_ nightly builds, i think. 15:34
* kparal is afk for 10 minutes 15:34
red_alert tflink: sure, but branching happens long after that afaik 15:35
adamw there's that place in the devel tree where a boot.iso gets built nightly, right? 15:35
tflink depending on the timing of all of this, the image building project could be an option 15:35
tflink but that's probably a discussion for another day since it isn't even close to being finished :) 15:36
nirik adamw: no, only in branched, rawhide produces no images IIRC. 15:36
adamw oh, yeah. it's just a repository. well, boo. 15:36
red_alert not anymore ever since nightlies are based on branches 15:36
adamw so yeah, good call red_alert...if anaconda gets into rawhide significantly before branch, we should ask releng to do nightlies if possible 15:37
adamw branching is scheduled for 07-31. 15:37
adamw nirik: how hard would it be to get a nightly boot.iso from rawhide for a few weeks? 15:38
nirik not too hard I wouldn't think... the branched script already does do them. 15:38
adamw OK, great. 15:39
adamw anything else anyone can think of? 15:39
j_dulaney Wait, what happened to Rawhide nightlies? 15:40
adamw we stopped getting them when Branched happened, apparently. 15:41
nirik http://fedoraproject.org/wiki/No_Frozen_Rawhide_Proposal 15:41
adamw OK, if no-one can think of anything else regarding newUI we can let clumens go, thanks chris! 15:41
clumens sure 15:41
clumens i'll go... back to work on new ui. 15:41
j_dulaney Go forth, my grasshopper 15:42
adamw OK, so in non-newUI-but-still-F18-planning topics... 15:42
red_alert do we plan to set up special newui tracker/blocker bugs? 15:43
adamw if we're going to do most of our testing after merge, which is what it sounds like, i don't really see a need 15:43
adamw it'll just be 'the anaconda in rawhide' 15:43
j_dulaney So, no more Rawhide nightlies, ever? I thought they restarted post-release 15:43
red_alert adamw: right 15:44
adamw so as long as clumens plans to merge about as early as testing is useful, that simplifies several things. 15:45
nirik under the current setup, rawhide never makes images. 15:45
nirik since about 2010 15:45
j_dulaney nirik: Then what was it that I was pulling that was labelled Rawhide Nightly? 15:45
adamw so, I have to go through the retrospective feedback and come up with action items, but that's just one for me unless anyone else has a burning desire to help 15:45
nirik j_dulaney: dunno. or are you talking about the live image nightlies? 15:46
j_dulaney nirik: +1 15:47
nirik those do follow rawhide. 15:47
adamw oh, hey, some #s. 15:47
adamw #info clumens plans to merge the new anaconda UI into Rawhide as soon as custom partitioning is implemented and broadly works. he estimates one month for that. 15:47
nirik but it's a seperate process. 15:47
j_dulaney Ah, Okay 15:47
adamw #info we don't believe QA can achieve much practical testing until that merge point, but we asked clumens to provide one or two public test images before then so testers can familiarize themselves with the new UI 15:48
* j_dulaney understands now 15:48
adamw #info key areas of installer that provide much of the testable complexity are repository configuration, storage configuration and bootloader configuration 15:48
adamw #info if newUI lands in anaconda significantly before F18 branch point, we will ask releng to do rawhide nightlies until branching 15:49
adamw #action adamw to review F17 retrospective and come up with an action plan 15:50
adamw we also have a bit of a problem lately with Housekeeping stuff; https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora16 for example 15:51
adamw that stuff sort of falls neatly between QA, bugzappers and the program manager, and consequently we lost track of it when poelcat stopped doing it 15:52
adamw so we need to co-ordinate with robyn and the new program manager when we have one, to catch up on it for f17 and f18. 15:52
j_dulaney +1 15:53
* j_dulaney can talk to rbergeron about it at SELF 15:53
tflink has it always been done by hand? 15:53
red_alert isn't that scripts that are run by rel-eng but no-one opened tickets to get it done 15:54
adamw it's a messy mish-mash. 15:54
adamw bits of it are partly automated, bits of it that are just mailshots aren't 15:55
tflink do the automated parts still work w/ the bz upgrade? 15:55
adamw one bit of it, for instance, if i remember correctly, the bugzilla EOL stuff requires someone to design a query that lists out all the EOL bugs in bugzilla, and send the query to the RH bugzilla maintainers who run some kind of script that closes them. 15:55
adamw that's a good question. 15:55
j_dulaney It does seem like a good candidate for scripting 15:56
j_dulaney adamw; Is said query done from the web ui? 15:57
nirik I think it's done by rhit bugzilla folks. 15:58
* nirik is tired of it not being done, offers to step up and try and get it done. 15:58
red_alert the query-url seems documented in the wiki and the process says open a bug against RH eng ops 15:58
red_alert only updating descriptions is done by hand, really 15:59
nirik red_alert: url? 15:59
adamw nirik: the query is supposed to be done by someone our side and sent to engops. 15:59
adamw anyhow 15:59
adamw point being, we need someone who wants to take care of this stuff, ideally asap. it's another one i'm happy to do as well, but if anyone else wants it, yell. 15:59
nirik yeah, ok. 15:59
* nirik is also happy to help see it get done, but would be happy for adamw to do it. ;) 16:00
nirik from my understanding we need: a) f14 notice we are going to close in N days/weeks, b) f15 notice we are going to close in a month, c) f14 close after ndays/weeks, d) f15 close at eol. e) rawhide rebase to f17... prossibly anytime. 16:02
red_alert nirik: urls are on the wiki page, one for rebase, one for eol 16:02
adamw nirik: that's the stuff from f16 16:02
adamw nirik: so actually we need to do the stuff for the f16 cycle, make a page for the f17 cycle and do all that too, and then make a page for the f18 cycle and stay on top of it this time. 16:02
nirik yep. 16:02
red_alert F15 was never finished, either 16:02
adamw fun for all the family! 16:03
adamw really? excellent. 16:03
* nirik was noting the things we are all behind on. Agreed we need to stay on top 16:03
red_alert wel...maybe it has and the wiki page is not up to date 16:03
nirik no, f14 bugs were never noticed or closed. 16:03
j_dulaney Hmm 16:04
j_dulaney That was weird 16:04
adamw f14 bugs being noticed and closed is part of f16 cycle, not f15 cycle. 16:06
adamw remember, EOL date is *two* distros behind. 16:06
adamw https://fedoraproject.org/wiki/BugZappers/HouseKeeping/Fedora16 talks about 'Fedora 14 EOL Closure'. 16:06
* nirik nods. I don't care which cycle, it wasn't done. ;) 16:06
adamw #action adamw to work with rbergeron to make sure housekeeping gets done for f16/f17/f18 16:09
adamw okay, we're kinda running over time so let's hold other f18 planning for future meetings, unless anyone has anything super-critical? 16:09
j_dulaney Huge-o lag 16:09
adamw yeah, i know, dunno why i'm lagging :/ 16:09
akshayvyas adamw: lagging where?? 16:10
adamw i am. here. occasionally. 16:10
adamw last call for important f18 planning topics... 16:10
adamw OK, moving on... 16:12
adamw #topic AutoQA update 16:12
adamw do we have any shiny autoqa news, tflink/kparal? 16:12
tflink nothing worthy of note from me this week 16:13
kparal no, still recovering from F17 16:13
akshayvyas kparal: :) 16:13
adamw okey dokey 16:14
adamw #info no AutoQA news, team still in F17 recovery 16:14
adamw #topic open floor 16:14
adamw alright, anything important missing? 16:14
adamw speak now or forever hold your peace. 16:15
adamw or, uh, until next week any how,. 16:15
adamw okay... 16:18
* adamw sets meeting fuse for 2 mins 16:18
adamw ssssssss 16:18
adamw #chair tflink kparal 16:18
zodbot Current chairs: adamw kparal tflink 16:18
tflink interesting timing on that 16:18
adamw  ? 16:18
tflink chairing kparal and I after the fuse is set 16:20
tflink or is that the lag you were talking about earlier 16:20
red_alert tflink: just so you can never complain not having been chair ;) 16:20
adamw i chaired you just in case i forget to end the meeting again =) 16:21
adamw speaking of which... 16:21
adamw #endmeeting 16:21

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!