15:01:02 #startmeeting Fedora QA meeting 15:01:02 Meeting started Mon Sep 23 15:01:02 2013 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:02 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01:02 #meetingname fedora-qa 15:01:02 The meeting name has been set to 'fedora-qa' 15:01:07 #topic Roll Call 15:01:13 #chair adamw 15:01:13 Current chairs: adamw tflink 15:01:22 ahoyhoy 15:01:25 meeting thief 15:01:27 :P 15:01:39 * pwhalen is here 15:01:45 if I would have known you were around, I would have waited :) 15:02:40 I can end the meeting so that you can start it, if that would make you feel better 15:04:06 is here 15:04:06 we'll be missing all the brno folks today 15:04:09 * roshi 15:04:45 ah, public holiday? 15:04:48 tflink: no, fine by me 15:04:59 team building stuff 15:05:07 hey 15:05:11 what's meeting today ? 15:05:57 horrible creatures 15:06:06 terrible, terrible things that should never see the light of day 15:06:22 pretty much, yeah :) 15:06:38 * jreznik is here 15:06:46 * adamw wonders if viking_ice is hiding in the woods near the team-building exercise in full viking war mode 15:07:14 that brings strange visuals to mind 15:07:29 no team building - this year only for devels and qa to make theirs life together better, nicer, sweeter :) 15:09:10 oh god, i hope somebody called 112 15:09:36 * adamw imagines CNN crew arriving on scene 15:09:47 "My god! Can somebody tell me what plane crashed here?" 15:09:55 "No plane crash. Only team building exercises." 15:10:19 reminds me of my old job when we had a kickball tournament 15:10:20 welp, let's see 15:10:24 qa collided with devels 15:10:34 a couple of people ended up in the hospital 15:10:48 and that was the last time we were allowed to have a kickball tournament :) 15:10:50 they were the lucky ones? :) 15:11:21 * roshi must have always worked at the wrong place. They never tried to build a team... :p 15:11:33 not really, some of them needed physical therapy and took months to recover fully 15:11:34 tflink: well, year ago we played football (our, european one) against some school trip guys and the amount of blood, when members of our team hit themselves... 15:11:55 welp, i guess we have enough people to roll through the agenda 15:13:11 #topic Previous meeting follow-up 15:13:24 "pschindl to send out karma/testing request for not-yet-stable updates in RC2" - well, one way or another we got past that 15:13:31 #info "pschindl to send out karma/testing request for not-yet-stable updates in RC2" - probably done, now stale 15:13:43 "tflink to make sure some cloud testing is getting done" - assume that was done too? 15:13:44 jreznik: so the moral of this story is nerds + sports == injury :) 15:13:59 yeah, I did some testing on the RC4 AMIs 15:14:10 #info "tflink to make sure some cloud testing is getting done" - done, tim did some testing on RC4 AMIs 15:14:21 "tflink to file blocker bug about DOA i386 AMI" - so i presume that was done too? 15:14:30 yeah, we should probably make a new list of stuff that needs karma 15:14:42 adamw: nope, it got fixed without the blocker bug 15:14:50 never did figure out where to file that 15:15:00 would be great to figure it out 15:15:05 it was fixed in the cloud ks 15:15:07 (where to file such cases) 15:15:29 oh kay 15:15:45 #info "tflink to file blocker bug about DOA i386 AMI" - not done, though the bug got fixed: still need to figure out where to file such bugs 15:15:55 where does the cloud ks live? 15:16:03 is it part of spin-kickstarts? 15:16:04 good question 15:16:15 as it went out of our radar a few times and there was no formal blocker decision... 15:16:38 looks like a separate repo 15:16:45 commit messages go out to cloud@ 15:16:55 yep, it's not in spin-kickstarts 15:17:02 well, wherever it is, it probably needs to be packaged for self-hosting purposes... 15:17:23 https://git.fedorahosted.org/cgit/cloud-kickstarts.git 15:17:24 #action tflink to talk to releng about where to file bugs in the cloud .ks and whether it needs to be packaged in the distro 15:18:39 * satellit_e joined late 15:19:06 "tflink to propose moving 'boot from optical media' criterion from alpha to beta" 15:19:26 i think you agreed to do that, right? 15:19:38 that happened, was waiting for some more +1s before moving to criteria wiki pages 15:19:40 it does leave us in the rather interesting position that we don't directly require the optical ISO to, er, boot at all in alpha 15:20:04 not unless it's properly sized 15:20:20 It's cheaper to test with USB drives. 15:20:35 tflink: true 15:20:52 who's going to burn alpha to a DL DVD or a BD-R, anyways 15:21:03 dl dvds aren't THAT hard to get a hold of, but yeah. 15:21:03 * satellit_e I still think we should issue 2 DVD's with different DE and basics so both were under DVD max size 15:21:07 and size is hard to get properly for alpha... 15:21:16 they're not hard to find but they're expensive 15:21:31 satellit: not really relevant to this discussion, though. it'd still be possible for one of them to go over. 15:21:46 not cost prohibitive, but I'd still question why anyone would do that with alpha 15:22:21 people do weird things 15:22:22 but hey 15:22:39 changes look good to me 15:22:42 i'd say fire away 15:23:30 #info "tflink to propose moving 'boot from optical media' criterion from alpha to beta" - proposal was posted: https://lists.fedoraproject.org/pipermail/test/2013-September/117799.html 15:23:48 any follow-up i'm missing, having been away? 15:24:08 not that I'm aware of, no 15:24:26 the app installer thing that we were being volunteered for went away 15:25:52 testing "optical media" 15:25:57 is very easy to do with a vm 15:26:20 drago01_: not the same thing 15:26:52 drago01_: we've come across corner cases where the image hasn't worked right when written to an actual disc in the past, and it's something we always want to test explicitly 15:27:02 tflink: well i'm bringing that up later anyhow 15:27:05 ok 15:27:12 #topic Fedora 20 Alpha final work and retrospective 15:27:24 #info Fedora 20 Alpha was approved and will go out tomorrow: nice job all 15:27:33 looks like i just get in the way, huh 15:28:03 so the obvious action item that we need to get done today is to get commonbugs up 15:28:09 #action adamw to get F20 common bugs page up and populated 15:28:13 if anyone wants to help write entries, please do 15:28:26 roshi: have you worked on commonbugs at all yet? 15:28:34 nope 15:28:42 or, if I have I didn't know it 15:28:45 roshi: maybe i'll rope you in to help with that 15:28:47 oh, you'd know it :P 15:28:51 haha 15:28:53 https://fedoraproject.org/wiki/Common_F19_bugs 15:28:54 works for me 15:28:56 i created the skeleton page 15:28:59 tflink: thanks 15:29:09 but didn't actually write anything meaningful in it :) 15:29:12 so, anyone think of anything else we need to check in on before tomorrow? 15:29:34 not really 15:31:06 guess we might want to double check everything's lined up to publish the arm and cloud images 15:31:21 though it'd really be fpm or fpl stuff i think :P 15:31:34 yeah, it looked like the "first class cloud" stuff was done with websites last I checked 15:31:47 but agreed that it would be worth looking into 15:31:59 website guys are working on the support, I expect it should be ok (trying to follow it up) 15:33:20 I'll write up the arm install instructions today, there is a place holder there that needs some edits 15:33:32 cool, thanks 15:33:36 #link https://fedorahosted.org/fedora-websites/ticket/229 15:33:44 #info looks like things are mostly in place for ARM and cloud image publication 15:34:09 adamw, if i come across something for common bugs, I'll also add to the page you mentioned 15:34:45 pwhalen: you should also mark it with the CommonBugs keyword and if you add it into the page, add the URL to Whiteboard 15:34:53 helps with tracking 15:35:05 adamw, okay, will do 15:35:07 this is explained in the comments to the commonbugs wiki source, which apparently no-one ever reads 15:35:18 :) 15:37:35 #info no-one can think of much else that needs doing for Alpha prep 15:37:48 #topic Fedora 20 Beta planning 15:38:08 * Viking-Ice is doing team building exercise 15:38:11 anything anyone can think of going forward for beta that we should consider now? any lessons from alpha cycle - criteria changes we need to do, test cases that needs fixing/adding? 15:38:18 the test matrix needs work . spins particularly https://fedoraproject.org/wiki/Test_Results:Fedora_20_Alpha_RC4_Desktop#Non_release-blocking_desktops:_x86_.2F_x86_64 15:38:29 has anyone tried out gnome+wayland? 15:38:40 also The test page link needs to be prominently displayed on the install page : https://fedoraproject.org/wiki/Releases/20/Spins as it is critical for the spin release being done. "yes" in right collums being criteria. 15:39:15 the test case problem from alpha (updates.img) was fixed, IIRC 15:39:41 yeah, pschindl created a new updates.img 15:39:45 tflink, I wanted to try ti because it shows up in gdm sessions, but later I heard gdm cannot start shell on wayland yet 15:39:51 * jreznik has to leave now... will be online later to track alpha readiness :) 15:40:06 satellit: when you say 'needs work', you mean we need to do the testing? or the matrix itself is busted? 15:40:14 tflink: not yet, probably a good idea 15:40:15 bitlord: ok, I thought it was working for the most part 15:40:24 tflink, it is possible to do it from terminal I think, but no command here, have it on another machine in IRC log from #gnome-shell @gimpnet 15:40:31 #info tflink suggests that we should do some gnome-on-wayland testing 15:40:32 the matrix is not very relevant for Soas,,, 15:40:58 http://wiki.sugarlabs.org/go/Fedora_19-sugar_test_cases is list I did of f19 15:41:17 for* 15:42:06 secureboot 15:42:10 tflink, I actually was hoping for the venue for gnome+wayland would be Gnome-SIG product so it would not be disruptive to the general desktop product to prevent us having to relive the graphics driver disruption all over again but but hey RH Desktop team successfully shot that down 15:42:38 Viking-Ice: it's only planned as a preview for F20 15:42:39 satellit: oh right, and i never got around to fiddling it into the matrix :( maybe you should do it, heh 15:43:02 ok will do so if it is oK 15:43:02 Viking-Ice: right, i think we just want to check it out so we know roughly where it's at and can answer questions 15:43:09 I just think it'll get a decent amount of attention if it mostly works and would be worth the time to poke at it and make sure it isn't completely busted 15:43:18 tflink, as an desktop product even as a preview it would have to pass our test matrix 15:43:23 satellit: if you edit it into the template page i can have a look over and let you know if it looks ok to me if you want 15:43:39 ok will work on it thanks 15:43:47 Viking-Ice: i don't see it that way...it's explicitly listed as a preview in f20 15:44:02 adamw, irrelevant release blocking desktop 15:44:10 hence one of the reason for the Gnome SIG 15:44:18 Viking-Ice: wayland isn't release blocking for f20, though 15:44:27 Viking-Ice: the fact that it's a release-blocking desktop doesn't mean we require absolutely every bit of code it contains to work. 15:44:48 i don't think anyone besides you is working on the basis that gnome-on-wayland is release blocking for f20...:) 15:44:54 i certainly haven't been 15:45:40 adamw, good for you 15:45:46 same here, I really doubt that it'll be ready for everyday usage for F20 15:45:55 no-one else seems to be under that impression either from what i've seen 15:47:03 bitlord: indeed, i didn't get the start wayland from GDM code ready in time for feature freeze 15:47:18 it's sitting on a branch, but unfinished 15:47:35 so as things stand we're scheduled to go straight for a blocker meeting on wednesday 15:47:53 do we have any proposed blockers? 15:47:54 adamw, well based on fesco dictating how they go about things I cannot see how no one does not see it that way 15:47:55 yeah 15:47:55 https://qa.fedoraproject.org/blockerbugs/milestone/20/beta/buglist 15:48:26 Anaconda really always that much 15:48:28 * tflink changes current to beta 15:48:57 Viking-Ice: i don't really see that anything written down anywhere says gnome-on-wayland would be release blocking. it's not absolutely explicitly specified in the criteria pages that it only applies to X, but then it doesn't state that it applies to all graphical servers either. 15:49:35 adamw, well we are always changing the criteria anyway so let's adjusted the invevitable now 15:49:59 Viking-Ice: we could explicitly specify it, i suppose, but there does have to be a common-sense line *somewhere* 15:50:26 adamw, that line we move every release cycle even during the testing phase so... 15:50:44 or else i'd have to write stuff like 'on the default display server, so long as a keyboard is plugged in, and you have enough memory, and you're not remoting your desktop via RFC 1149...' 15:51:16 it's not really practical to encode EVERY goddamn assumption or you wind up writing eu food safety legislation :P 15:51:26 well hopefully I manage to kill the gnome tunnel vision in the project and we in QA start focusing strictly on the installer and the core/baseOS 15:51:36 that's really kind of a separate question 15:51:37 anyhow 15:51:48 sounds like blocker meeting on wed makes sense 15:51:58 #info first Beta blocker meeting will go ahead on Wednesday 09-25 15:52:22 yeah, makes sense to have one - already have several proposed blockers 15:52:30 * adamw thinks any day you can work a reference to RFC 1149 into the conversation is a good day 15:52:53 TC1 is scheduled for 10-01, next tuesday 15:52:53 * tflink makes mental note to avoid conversations with adamw 15:52:55 certainly early enough 15:53:05 halfline, ok, but would be nice to have somewhere in documentation how to test wayland on fedora (as technology preview, or how they call those things which aren't features ;-) 15:53:07 go/no-go is 10-24 15:53:18 #info TC1 scheduled for 10-01 (next Tuesday), go/no-go for 10-24 15:53:46 .fire tflink 15:53:46 adamw fires tflink 15:53:53 insubordination! 15:54:11 bitlord: yea makes sense 15:54:24 halfline, and probably "hide" it from gdm sessions, so people don't get confused when trying and it doesn't work, so no unneeded bugs reported 15:54:26 though i'm hoping to get a f20 update out that enables testing when it's ready 15:54:32 so any known criteria alteration we need to do before we start the beta process 15:55:09 halfline, /me not a qa member, just do some testing sometimes ;-) 15:55:17 i need to put the update criteria change into production i guess 15:55:26 bitlord: that pretty much makes you a qa member :P 15:55:33 adamw, as soon as yesterday 15:55:35 ;) 15:55:38 adamw, hehe, thanks ;-) 15:55:39 #action adamw to send the update criteria to 'production' 15:55:42 adamw: you forgot to put the reason into the .fire :) 15:55:46 in retrospect the wayland session file probably shouldn't be in /usr/share/xsessions 15:55:59 .fire tflink again: nitpicking 15:56:00 adamw fires tflink again: nitpicking 15:56:04 since it's not an X session, and since display managers need special handling to support it 15:56:24 halfline: yeah, there are tools/docs which assume anything there is a perfectly normal X session, i think. 15:58:07 any other criteria changes required? i can't think of any 15:58:20 * adamw should check if anything needs changing for arm/cloud 15:58:26 requiring optical boot @ beta 15:58:29 but we already covered that 15:58:32 #action adamw to check if any further criteria changes are needed to beta/final for arm/cloud 15:58:46 tflink: it already *does*, the only necessary change is to the 'explanatory text' 15:58:54 but that's not critical 16:00:24 adamw: okay i got gnome release-team approval to change that before 3.10 today 16:01:33 cool 16:01:52 okay then, we're over time, so a quick: 16:01:54 #topic Open floor 16:01:56 so the cloud testing for beta is it solid this time or will we be experiencing those hickups as we did during alpha 16:01:58 any other business? 16:02:12 Viking-Ice: depends on if we do the testing, I imagine 16:02:27 Viking-Ice: well i asked earlier if there was anything we needed to change based on arm/cloud experiences for alpha and no-one had any suggestions... 16:02:35 tflink, well it's up to the cloud community to do the testing 16:02:43 in theory, yes 16:02:48 in practice, though? 16:02:51 yes 16:03:11 I don't remember them ever helping out with testing unless someone specifically asked for it on cloud@ 16:03:27 sub community in general should do their testing 16:03:27 s/ever/ever helping out with much/ 16:03:40 sure, but you also used the "s word" 16:03:50 well, maybe we need a better process for pulling them into testing 16:03:58 and I thought robin was working on granting access to people 16:04:19 I really don't think access was the problem 16:04:31 at full price, an hour with a micro instance is USD 0.20 16:04:38 tflink: your substitution skills suck 16:04:47 s/0.20/0.02 16:04:55 tflink: plus the barrier of 'i have to give amazon my visa card so i can test this thing?' 16:04:56 adamw: or I'm just lazy :) 16:05:17 true, I already have it set up for other things 16:05:31 why on earth did we approve something to be release blocking while it has a prize tag on it 16:05:49 amazon isn't the only way to test for criterion violation 16:05:52 it's just the easiest 16:06:03 the way that the crition is worded, it just needs to work with xen 16:06:22 tflink, dont we already have virtualsation criteria that covers xen ( as well as kvm ) 16:06:39 Viking-Ice: there is a separate cloud requrement 16:06:59 tflink, is that due to adamw "Recommending kvm" 16:07:07 unrelated 16:07:15 xen DomO is not a release blocking issue 16:07:29 we should just have one criteria that covers "all virtual" 16:07:37 no, we shouldn't 16:07:38 Viking-Ice: then we're on the hook for vbox, which we don't want to be. 16:07:55 and we want to support only domU for xen, but both host and guest for KVM. 16:07:55 vbox, vmware, hyperv ... 16:08:01 er, 'block on', not 'support'. 16:08:04 adamw, arent we playing that game with amazon 16:08:11 playing what game? 16:08:20 the most common virtualization for cloud is kvm and xen 16:08:24 the support 16:08:33 Viking-Ice: EC2 is xen. 16:08:38 so is rackspace 16:08:56 and so our virtualsation should cover xen/kvm 16:09:01 we agreed EC2 would be a 'blocking platform' for f20 cloud stuff because, basically, the world and their mom runs stuff in ec2 16:09:06 and infra should set up host internally to test that stuff 16:09:10 it's kind of a practical choice 16:09:25 Viking-Ice: it does 16:09:32 not really far from it actually practical for rhel I suppose 16:09:36 Viking-Ice: and they are. it's just that it'd be nice to check everything works as intended in the exact amazon ec2 configuration so you can just click the AMI or whatever and be done 16:09:51 that's one of the main goals of the whole 'first class cloud' thing 16:10:30 I dont mind first class cloud thing as we host and test it privately as in our infrastructure on a fedora host 16:10:54 bring it up with fesco 16:11:04 lol no thank you sure 16:11:06 mean sir 16:11:27 that's where the first class cloud stuff came from 16:11:45 well, the feature process 16:11:50 ah I see that explains it 16:12:01 it is a feature for f20, IIRC 16:12:18 Viking-Ice: it's kind of a 'map is not the territory' thing 16:12:27 bottom line gnome sig is also result from fesco decision making 16:12:31 what we're trying to ensure is 'you can just click a button and up pops f20 in ec2' 16:12:37 the only way you can really test that reliably is *to test it in ec2* 16:12:45 we can never guarantee that we have a perfect internal copy of ec2 or something 16:12:59 which brings us back to what the cloud sig should be doing 16:13:09 and using the s-word again 16:13:16 but making it release blocking is arguably questionable 16:13:23 like i said, perhaps the thing we should be aiming for is to have a more robust process for getting the cloud sig involved in validation testing 16:13:34 did we do anything specific down that line for alpha? or just trust that they'd follow along? 16:13:53 well we did ping them ( and them being matt ) 16:14:01 they said they did testing, and we believed it until I found the x32 image to be DOA 16:14:41 we can try pestering them on cloud@ more 16:15:23 in the end of the day this is their "product" and their responsibility and if it fails to shine on amazon then it's their fault 16:15:29 to be blunt 16:15:35 sure but we still look bad 16:15:43 no they look bad 16:16:01 the only time we look bad as a project is when we have "defaults" that look bad 16:16:10 anyhow 16:16:11 because your average ec2 user isn't going to say "fedora sucks on the cloud, they can't test this stuff" 16:16:17 we're 16 minutes over at this point 16:16:20 i think we got the juice out 16:16:24 probably not, but we're getting into flamewar territory :) 16:16:30 yep 16:16:46 #info we should look at ways to tie cloud SIG into validation testing more closely and verify their work 16:17:02 were they actually filing bogus entries in the matrices? or were they just saying 'yeah sure, we tested that'? 16:17:10 the latter 16:17:13 well we see who get's the blame when Gnome removes the middle mouse click paste Fedora or Gnome ( default or not ) 16:17:40 i didn't think that was on purpose, was it? 16:18:03 it was on purpose, yeah. just a little premature. 16:18:07 tflink, it's going to be 16:18:08 but i'd say it'd be gnome that gets the blame :P 16:18:15 buty we're definitely off topic now 16:18:18 * adamw sets quantum fuse 16:18:30 I'd ask why they removed it, but I know better 16:19:47 tflink, I spent 1 and half our watching the designer and developers arguing before I gave up. it felt like watching highlander 2 all over again and literally knew I those minutes of my life I would not get back so yeah one know better 16:20:16 tflink: i think the idea is to use middle-click for something shinier and more modern, but my eyes kinda glazed over at some point. 16:20:24 Viking-Ice: yeah, that sounds like what I expected 16:21:02 * roshi didn't even know his scroll wheel button *ever* did anything useful 16:21:36 and you have been using linux for how long again? 16:21:38 how did I not know this? It's useful 16:21:49 roshi: wait, you didn't know that? 16:21:53 well - I try to not use a mouse if I can help it 16:21:55 .fire roshi insufficient geekiness 16:21:57 adamw fires roshi insufficient geekiness 16:22:03 and there it is! 16:22:05 alright, thanks for coming folks 16:22:10 #endmeeting