16:00:17 <adamw> #startmeeting Fedora QA meeting
16:00:17 <zodbot> Meeting started Mon Jan 26 16:00:17 2015 UTC.  The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:17 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:20 <adamw> #meetingname fedora-qa
16:00:20 <zodbot> The meeting name has been set to 'fedora-qa'
16:00:25 <adamw> #topic Roll call
16:00:27 * roshi_ is here
16:00:36 <adamw> ahoyhoy folks
16:00:43 <adamw> who's here to A the heck out of some Q?
16:01:13 * pschindl is here
16:01:13 * mkrizek is here
16:02:16 * danofsatx is heah
16:02:36 <adamw> thanks for coming out
16:02:42 <adamw> #chair roshi danofsatx
16:02:42 <zodbot> Current chairs: adamw danofsatx roshi
16:03:02 <adamw> pschindl: kparal out today?
16:03:16 * kparal here
16:03:25 <adamw> aha, lurking away
16:03:28 * satellit listening
16:05:07 * nirik is lurking, but also going to go get more coffee.
16:05:24 <adamw> #topic Fedora 22 testing status
16:05:43 <adamw> so if you can believe it, we're getting towards Alpha TC1 time now...
16:05:52 <roshi> woohoo!
16:06:03 <adamw> inorite
16:06:15 <adamw> I mostly put this topic in to explain why we weren't getting nightlies, in fact
16:06:20 <adamw> but then we got one, so yay?
16:06:25 <roshi> yea
16:06:26 <kparal> yay
16:06:38 <satellit> some are working
16:06:59 <adamw> yeah, I suspect we've still got quite a few bugs to find in the dnf stuff though
16:07:13 <satellit> boot.iso seems to work now
16:07:15 <adamw> we should try and exercise that with different types of package selection, kickstart repo directives and stuff
16:07:32 <roshi> sounds like a good idea
16:07:42 <adamw> looking at https://www.happyassassin.net/testcase_stats/22/ , we've definitely covered some good ground but there are a lot of tests still to do
16:08:30 <danofsatx> looks like I need to get on the ball.
16:08:37 <adamw> nirik: should the nightly train be back on schedule now? or was it a one-day fix?
16:08:50 * danofsatx checks his daily calendar to carve out 3 hours between 2 and 5 a.m.
16:09:21 <Corey84> .fas corey84
16:09:22 <zodbot> Corey84: corey84 'Corey84' <sheldon.corey@gmail.com>
16:10:32 <adamw> danofsatx: remember to sleep, we don't need qa zombies ;)
16:10:35 <adamw> ahoy corey
16:10:47 <danofsatx> I have sleep scheduled from 11pm to 2am.
16:10:57 <adamw> #topic Fedora 22 Change check-in
16:11:10 <adamw> danofsatx: three whole hours?! that's 2 hours you could use for testing right there
16:11:16 <adamw> ;)
16:11:41 <adamw> so the F22 Change list is filling out: https://fedoraproject.org/wiki/Releases/22/ChangeSet
16:11:43 <danofsatx> I know, but I'm getting too old to run on coffee alone.
16:12:06 <adamw> there are still more proposed Changes that FESCo has to review, at https://fedoraproject.org/wiki/Category:ChangeAnnounced
16:12:51 <adamw> are there any in particular that have caught anyone's eye yet?
16:13:05 <adamw> stuff that's going to need significant testing and could have a major impact on release quality
16:13:25 <kparal> python 3 can cause some issues in anaconda
16:13:38 <kparal> strings handling
16:14:00 <kparal> if they intend to run it on python 3, haven't checked
16:14:00 <adamw> yeah, that's one of the obvious ones
16:14:04 <adamw> it's an open question
16:14:09 <adamw> https://fedorahosted.org/fesco/ticket/1392
16:14:12 <danofsatx> ah crud, the page I read last week in prep was the accepted, not announced
16:14:21 <adamw> #info Not yet clear if anaconda is intended to be ported to python3 for F22: https://fedorahosted.org/fesco/ticket/1392
16:14:36 <kparal> Preupgrade Assistant? huh, what's that?
16:14:38 <adamw> danofsatx: accepted is *later* than announced (bit confusing i know)
16:14:55 <danofsatx> yeah, I got that (now)
16:15:13 <danofsatx> announced should be renamed to 'proposed'
16:15:40 <adamw> there are multiple stages of 'proposed'
16:15:49 <adamw> there's also ReadyForWrangler but i didn't link that as we're past that stage now
16:15:56 <danofsatx> ok, 'submitted for consideration'
16:16:12 <kparal> that Preupgrade Assistant might be of interest to us
16:16:25 <kparal> and it's a bit unfortunate name, imho
16:16:51 <satellit> afk
16:18:28 <adamw> it'll be worth testing, but it doesn't seem like super critical stuff as it's a sort of optional add-on to fedup
16:18:36 <adamw> it'll depend a bit on how heavily it gets promoted i suppose
16:18:54 <adamw> if all the upgrade instructions tell people to use it, obviously it's important...we could ask about that?
16:19:56 <adamw> kparal: did you want to take an action item to follow up on it?
16:20:57 <kparal> hmm, yes, sure, that's definitely what I wanted
16:21:01 <kparal> :)
16:21:25 <kparal> I'll talk to the maintainer
16:21:33 <adamw> ;)
16:21:42 <kparal> he's probably from Brno anyway
16:21:52 <adamw> #action kparal to check in on Preupgrade Assistant feature, what testing is desired and how heavily it will be promoted
16:22:05 <adamw> i agree the name is a bit unfortunate too, i wonder if anyone at fesco noticed that...
16:22:30 <nirik> adamw: so, we downgraded mock to get nightlies working... we now think we have a fix for the new one, which we need to test and then deploy. Hopefully they will keep working...
16:22:33 <adamw> (in case anyone's not sure what we mean, the predecessor tool to fedup was called 'preupgrade')
16:22:43 <adamw> nirik: roger, thanks
16:24:03 <adamw> so the other one that catches my eye obviously is yum->dnmf
16:24:05 <adamw> dnf*
16:24:19 <adamw> there's been a big devel@ list discussion about that one
16:24:44 <adamw> did anyone have any particular thoughts on the topic? worries about whether dnf is ready, ideas about how to do rigorous testing, etc
16:25:13 <roshi> I don't have anything specific
16:25:24 <danofsatx> I'm not worried that dnf isn't ready, I'm quite sure that it isn't.
16:25:34 <adamw> heh.
16:25:49 <adamw> danofsatx: do you have particular shortcomings in mind? are they filed as bugs?
16:25:56 <adamw> maybe i shouls say 'won't be ready by final'
16:26:07 <roshi> I just worry how much testing it's going to need, and if bugs'll get fixed fast enough
16:26:21 <danofsatx> anecdotal evidence through @lists and here on IRC. I haven't personally devoted any time to it as of yet.
16:26:34 <Corey84> ^ same
16:27:23 <danofsatx> I'll start using it and documenting my findings - I only have a few straggling F20 systems around, but I'll deploy a couple rawhide VMs to start in depth hammering.
16:28:04 <adamw> part of the issue is the change doesn't only encompass users using it interactively
16:28:54 <adamw> anaconda's switched to it already, and the change describes moving release engineering tasks to it
16:29:19 * nirik doubts those will happen/be possible.
16:29:39 <adamw> the change page doesn't mention switching fedup, i should probably check on that
16:30:06 <roshi> that's an ambitious change
16:30:08 <danofsatx> action item?
16:30:22 <Corey84> maybe i missed it  but was anaconda gonna go to py3?
16:30:36 <adamw> is anyone interested in sort of taking charge of QA for this change? trying to come up with a plan for testing beyond just 'have a bunch of people use dnf and hope for the best'?
16:30:40 <danofsatx> that's under consideration by the anaconda folks, Corey84
16:30:42 <adamw> Corey84: it's not decided yet
16:30:47 <adamw> Corey84: https://fedorahosted.org/fesco/ticket/1392
16:30:48 <roshi> in one release we're going to switch everything from yum to dnf and everything from python2 to python3
16:30:48 <Corey84> kk
16:31:06 <Corey84> that is ambitious indeed roshi
16:31:08 <danofsatx> well, didn't we switch everything to systemd in one release?
16:31:11 <adamw> roshi: the python3 feature doesn't involve changing 'everything' to python3, but yeah, they're two major changes
16:31:26 <adamw> in a short release cycle
16:31:44 <roshi> I know - was being faceitous "Hey, while we're here, lets CHANGE ALL THE THINGS!"
16:32:02 <danofsatx> while we're at it, lets rename the project.
16:32:06 <Corey84> adamw,  woulden't it only require  non py3 stuff to be back/forward compat ?
16:32:29 <Corey84> lets stop the non sense :P
16:32:50 <Corey84> this is fedora afterall right not ubuntu
16:33:05 <adamw> Corey84: the python3 change is about defaults and major fedora components, python2 is still included in fedora and things can still use it
16:33:12 <roshi> I would take the dnf action item, but I'm not that familiar with it
16:33:53 <danofsatx> roshi - no need to. adam said " i should probably check on that"
16:34:19 <roshi> he did?
16:34:21 <Corey84> roshi,  im willing to take that with you as a double team kinda thing
16:34:37 <Corey84> or adamw rather
16:34:54 <danofsatx> in regards to dnf and fedup, anyhow. the testing phase, well....that does need an action item.
16:35:07 <Corey84> need to wrap my head around that more anyways
16:35:22 <roshi> I thnk that's what adam was talking about, mapping out what all DNF would need tested
16:35:32 <roshi> it's what I was talking about anyways
16:36:20 <roshi> that'd work for me Corey84 - unless adam is already on top of it
16:36:49 <Corey84> regardless I'm game to help already slowly add stuff to my plate anyway
16:37:08 <adamw> roshi: right, if you want to take that it'd be great
16:37:17 <adamw> i don't think any of us have detailed pre-existing knowledge, so it's the same for anyone
16:37:24 <roshi> psh
16:37:35 <roshi> adamw == detailed pre-existing knowledge
16:37:38 <roshi> we all know that
16:37:40 <roshi> :p
16:37:42 * roshi ducks
16:37:48 <danofsatx> he wrote it, didn't he?
16:38:18 <adamw> roshi: it'd be a good idea to check that devel@ thread and see what got brought up there; ideally i'd like it if we got some really large-scale scripted testing in to compare yum's and dnf's behaviour in scenarios like 'installing each package in the repos on a clean minimal install' and stuff
16:38:29 <roshi> yeah
16:38:33 <roshi> that'd be good
16:38:38 <adamw> then make it so ;)
16:38:51 <roshi> I'll dig through the thread - I haven't been following it really
16:39:12 <adamw> #action roshi and corey84 to try and work up some methodical testing plans for dnf
16:39:46 <adamw> basically we want as far as possible to methodically/rigorously compare the real-world behaviours of yum and dnf and really find out where they're different
16:39:52 <adamw> yell for help if you need it
16:40:28 <adamw> there's also a systemd package split Change in the proposed list which could affect some scenarios, but that's probably the kind of thing where we should catch any major breakage it causes in regular testing
16:40:31 <adamw> just one to be aware of
16:40:42 <roshi> sure
16:40:58 <roshi> I'll walk outside and do a test yell, let me know if anyone can hear my screams
16:41:22 * adamw closes and locks the door, dusts off hands, proceeds
16:41:32 <adamw> #topic Release criteria status
16:41:38 <roshi> lol
16:41:46 <adamw> goddamnit it's freaking Houdini over here!
16:42:46 <roshi> :)
16:43:24 <adamw> so i think i need to rework that package set criterion *again* as apparently we're getting per-flavor (that's the new word for products, folks, learn to love it!) netinsts for f22
16:43:31 <adamw> plus a generic one
16:43:45 <adamw> so i'll work out how to change it for that, and send a new proposal
16:43:49 <roshi> yeah...
16:43:52 <roshi> sgtm
16:44:09 <adamw> then we need to do the multiboot criteria, thanks to cmurf for feedback on that, need to make sure pjones is ok with whatever we pick
16:44:32 <adamw> #action adamw to revise the package set criterion proposal again and move forward with multiboot proposal
16:44:47 <adamw> did we have any others that need attention that i've forgotten about?
16:44:59 <roshi> I don't think so?
16:45:09 * roshi has been in dev mode
16:45:11 <adamw> oh, i have a proposal in with server@ to move the role requirements out into separate pages
16:45:13 * kparal just notes that secure boot is not working with win8 booting now, if that needs to be taken into account when reworking those criteria
16:45:35 <adamw> kparal: yeah, i think i've got the bug report for that one, thanks...
16:45:52 <kparal> but allegedly it has been working with opensuse for years, so we might also want to do something about it, perhaps
16:45:55 <adamw> https://lists.fedoraproject.org/pipermail/server/2015-January/001719.html is the server role requirement split proposal
16:46:10 <kparal> of course it's easy for me to say that, when somebody else does all the work
16:46:14 <adamw> kparal: i read that too, sounded odd, but i've not looked into it in any detail
16:47:09 <adamw> #info adamw has a proposal to move the specific requirements for each server role out of the criteria and into role-specific pages - https://lists.fedoraproject.org/pipermail/server/2015-January/001719.html
16:48:20 <adamw> alrighty, i guess that's that for now
16:48:20 <adamw> #topic Open floor
16:48:33 <adamw> remember folks, we have blocker review coming up after this meeting, don't go away (just go to #fedora-blocker-review )
16:48:34 <tflink> as a heads up, it sounds like the current plan for bodhi2 is to switch over during the f22 cycle
16:48:35 <roshi> Blocker review after this meeting in #fedora-blocker-review
16:48:52 <adamw> tflink: good lord, are we riding there on our flying unicorn pigs?!
16:48:53 <roshi> I should read before hitting enter
16:48:57 <Corey84> kparal, secure + efi ?
16:49:09 <kparal> Corey84: the question is?
16:49:15 <adamw> Corey84: if you install fedora on top of Windows with secure boot installed, the entry in fedora's boot menu to boot Windows won't work
16:49:19 <Corey84> have noticed some edge case failure when using both even on 21
16:49:27 <adamw> Corey84: you can still boot Windows from the UEFI menu
16:49:38 <adamw> (aiui, anyway)
16:49:40 <tflink> adamw: not sure, seems like a questionable plan to me but I've not been following bodhi2 much
16:49:58 <Corey84> not what i was thinking was thinking the other way around
16:50:52 <pjones> So I'm okay with taking the shim patch upstream and including it next time we're rebuilding/re-signing shim anyway, but I'd rather not do a build/sign cycle just for this.
16:51:07 <roshi> I wish we were doing release names still
16:51:16 <roshi> I'd call F22 "Multivariate"
16:51:24 <pjones> especially since you can still boot into the other OS with firmware boot menus or efibootmgr -n
16:52:06 <pjones> (which I think is why we were getting mixed results before, probably - not realizing it wasn't going through grub in some testing.)
16:52:40 <kparal> some uefi don't have one time boot menu, I heard. in that case it's uncomfortable
16:52:50 <pjones> hence efibootmgr -n
16:52:51 <adamw> pjones: 'the shim patch upstream' is what makes this work for suse?
16:52:54 <kparal> pjones: but thanks for the update, great that you know about it
16:52:57 <pjones> adamw: yeah
16:52:59 <adamw> k
16:53:19 <adamw> pjones: are we likely to get a new shim build during f22?
16:53:22 <pjones> I mean I still need to, you know, actually review it and all.
16:53:26 <adamw> pjones: pshaw
16:53:27 <pjones> uh... maybe.  unclear.
16:53:29 <adamw> ok
16:54:05 <adamw> i guess we can add a commonbugs and document 'efibootmgr -n', i wasn't aware of that one
16:54:28 <kparal> at least we know we shouldn't put it into criteria just yet
16:55:20 <adamw> #action adamw to document 'efibootmgr -n' workaround for failure to boot windows from grub with SB enabled (#1170245)
16:55:42 <danofsatx> roshi: F22 Raptor. it's already been decided.
16:56:14 <adamw> F22 Velociraptor
16:56:44 <roshi> F22 Multiraptor
16:56:59 <tflink> that sounds like a multi-headed raptor
16:57:09 <tflink> or multiple raptors duct taped together
16:57:11 <roshi> of course
16:57:15 <tflink> not sure which scares me more
16:57:25 * adamw notices randall munroe in a foetal position in the corner, wibbling
16:57:27 <roshi> raptors are like voltron, the more you hook up the better it gets
16:57:43 <danofsatx> f22: http://goo.gl/lqOwIe
16:59:11 <adamw> i think the fedora version looks a bit more like http://st.pixanews.com/wp-content/uploads/2011/08/722.jpg
16:59:19 <adamw> ...and on that note, i think it's time for blocker review ;)
16:59:22 * adamw sets a short fuse
16:59:26 <adamw> thanks for coming, folks!
16:59:46 <roshi> now onto the fun part :)
16:59:53 <roshi> thanks for hosting adamw!
17:00:02 <adamw> #endmeeting