17:00:26 #startmeeting F30 Beta Go/No-Go meeting 17:00:26 Meeting started Thu Mar 21 17:00:26 2019 UTC. 17:00:26 This meeting is logged and archived in a public location. 17:00:26 The chair is bcotton. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:26 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:26 The meeting name has been set to 'f30_beta_go/no-go_meeting' 17:00:31 #meetingname F30-Beta-Go_No_Go-meeting 17:00:31 The meeting name has been set to 'f30-beta-go_no_go-meeting' 17:00:36 #topic Roll Call 17:00:40 .hello mohanboddu 17:00:41 mboddu: mohanboddu 'Mohan Boddu' 17:00:44 .hello2 17:00:47 frantisekz: frantisekz 'František Zatloukal' 17:00:52 * satellit listening 17:01:16 good $daypart to you mboddu, frantisekz, satellit 17:01:27 bcotton++ 17:01:31 frantisekz: Karma for bcotton changed to 16 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:01:32 morning 17:01:46 evening 17:02:01 hi nirik 17:02:20 we'll give folks a few more minutes to wander in 17:02:21 ain't no party like a go/nogo party. 17:02:23 .hello2 17:02:24 bowlofeggs: bowlofeggs 'Randy Barlow' 17:02:37 .hello adamwill 17:02:38 adamw: adamwill 'Adam Williamson' 17:02:49 .hello pwhalen 17:02:50 pwhalen: pwhalen 'Paul Whalen' 17:02:50 ...'cause a go/nogo party don't stop (until we have made a decision) 17:02:56 .hello2 17:02:57 coremodule: coremodule 'Geoffrey Marr' 17:02:58 hi bowlofeggs, adamw, pwhalen, coremodule 17:03:13 what happens if someone calls the cops on our go/nogo party? 17:03:16 hi bcotton 17:03:23 depends if they're go cops or nogo cops 17:03:57 adamw++ 17:04:23 okay, looks like FESCo, RelEng, and QA are all well-represented, so let's get going 17:04:29 .hello2 17:04:30 sgallagh: sgallagh 'Stephen Gallagher' 17:04:34 #topic Purpose of this meeting 17:04:35 adamw: is QA the go/nogo cops? 17:04:35 #info Purpose of this meeting is to check whether or not F30 Beta is ready for shipment, according to the release criteria. 17:04:37 #info This is determined in a few ways: 17:04:43 #info 1. No remaining blocker bugs 17:04:45 #info 2. Release candidate compose is available 17:04:47 #info 3. Test matrices for Beta are fully completed 17:04:54 were there go/no-go cops inside all of us, all along? 17:05:05 #topic Current status — Blocker bugs 17:05:20 bowlofeggs: maybe the real go/no-go cops are the friends we made along the way 17:05:25 haha 17:05:28 #link https://qa.fedoraproject.org/blockerbugs/milestone/30/beta/buglist 17:05:44 #info 0 Proposed Blockers 17:05:48 .fire bcotton 17:05:50 adamw fires bcotton 17:05:53 (^^ a Christmas miracle?) 17:05:55 bcotton: That's not true. 17:06:13 Oh, wait. I guess it's no longer "proposed" 17:06:17 #info 2 Accepted Blockers 17:06:48 so let's review this real quick and see if they're still real 17:06:51 #topic (1690566) Arm desktops do not have a browser 17:06:53 #link https://bugzilla.redhat.com/show_bug.cgi?id=1690566 17:06:55 #info Accepted Blocker, spin-kickstarts, POST 17:07:33 sadly this one is still there... 17:07:36 So, this is fixed in rawhide but not in the rc compose 17:07:48 #info fix was merged only to master and not the f30 branch so this is *not* fixed in Beta-1.4 17:07:55 When I merged the PR, I thought it was made for f30, but it was actually merged in master 17:08:26 * mboddu bangs his head 17:08:26 i've asked for a 1.5 build with this fix in it 17:08:30 that would be the only change from 1.4 17:08:39 adamw: that answers the question i was typign 17:08:40 so we could potentially go with that tomorrow on smoke testing 17:09:11 #info Beta-1.5 compose will only differ from Beta-1.4 by the fix for this bug 17:09:40 I assume we can't respin just the arm images? 17:09:52 frantisekz: Nope 17:09:55 Its all or nothing 17:10:05 anything else we want to say on this one? 17:11:20 #topic (1683197) gdm Fails to load with "nomodeset" 17:11:21 #link https://bugzilla.redhat.com/show_bug.cgi?id=1683197 17:11:23 #info Accepted Blocker, xorg-x11-drv-qxl, ASSIGNED 17:11:40 so, this one is awkward 17:11:54 we got a sort-of fix for it, only it turns out it only fixes UEFI. 17:12:07 i asked ajax for more info on how we can fix the BIOS case, but did not hear back yet. 17:12:22 it is definitely worth noting that F29 *final* contains this exact same bug. 17:12:37 fun. how did we miss it there? 17:12:42 though that was clearly a process fail on our part; as near as I can tell, this specific bug was never considered as an f29 blocker 17:12:43 we didn't 17:13:03 I think we've heard it's going to work when it needs to 17:13:10 but, it doesn't seem so 17:13:12 that was a different bug 17:13:32 i am curious as to how lruzicka and sumantrom claimed a pass for this test for F29 Final validation as i cannot see how it could possibly have worked 17:13:35 does f29 with updates have this bug? 17:13:35 anyhow... so now this causes nomodeset bios boots to not work? 17:13:37 i'll have to triple check it, but 17:13:55 i.e. would it be a regression for stable f29 users? 17:14:16 yeah, i believe the state for beta-1.4 is bios won't work, uefi may. 17:14:19 * nirik has no sense how many folks still use bios boot... 17:14:26 bowlofeggs: i think so. nothing has been done to fix it. 17:14:28 nirik: i think quite a lot. 17:14:31 adamw, uefi didn't work for me today on bare metal 17:14:42 my impression is that a lot of people have got 'disable UEFI and secure boot' as a meme 17:14:57 possibly so 17:14:59 frantisekz: there seem to be other bugs that can occur after this one to break it which *are* hardware-specific 17:15:09 my bare metal box doesn't work either, but it's clearly hitting a different bug 17:15:46 hmm, anyway, I'd incline to not block beta on this one, downgrade it to final blocker 17:15:53 *not to 17:16:48 agreed. my inclination is to say "forget the criteria! let's move ahead". but i don't know what that does to my opinion of myself 17:17:25 if a lot of users don't use UEFI, they might not be able to test the beta i'd guess 17:17:41 it may be more fair to say "i want the group to decide that, but i don't want to be the one to stand up and propose it" :-) 17:17:50 bowlofeggs: well, only if they need/select basic graphics 17:17:57 bowlofeggs: Well, it's the set of users that both don't use UEFI *and* can't use a mode-setting driver. 17:17:57 bowlofeggs, I don't think a ton of users use basic graphics 17:17:58 bowlofeggs: IF they only set nomodeset 17:18:04 That's... likely a very small set of users 17:18:06 oh ok 17:18:14 heh, firehose of replies! 17:18:28 i think there's at least a reasonable case to reconsider this criterion 17:18:36 but at the same time be aware that people *do* still at least try and use the option 17:18:47 adamw: are the other uefi bugs blockers/proposed? or ? 17:18:51 if you google 'fedora 29' 'nomodeset' or 'fedora 29' 'basic graphics' you will find results 17:18:56 nirik: not yet, we hadn't really got to that point 17:19:01 if we think it's not that many users, it wouldn't bother me to say it's a final requirement 17:19:06 Frankly, I think we should consider disabling this option in the bootloader for F31. 17:19:14 (and arguably they wouldn't be taken since so far they're all on hardware that works OK with regular graphics) 17:19:38 sgallagh: i nominate you to remember how to disable it everywhere. :P 17:19:45 (there are a surprising amount of places, iirc.) 17:19:52 adamw: I didn't say "disable it everywhere" 17:19:58 Just remove the bootloader entry 17:20:00 find and replace everywhere, what could go wrong 17:20:15 sgallagh: we have at least two bootloaders and i think more than that number of configs. but anyway, that's a sidebar. 17:20:22 right, back on target 17:20:45 Is anyone here prepared to die on this particular hill, or shall we agree to kick it to a Final blocker? 17:20:48 i would say i'm opposed to dropping this as a blocker *unless we agree to remove or move the criterion* 17:20:54 we can't just kick the bug 17:20:57 ok 17:21:11 I guess I am inclined to demote this to final, but the fact that there are other uefi boot issues makes me think we don't have enough testing time. ;) 17:21:14 as the criteria stand, and since people have consistently wanted to interpret the text as meaning 'the option should actually work at least a bit', it's a clear violation 17:21:30 so, scenario: 17:21:36 nirik: Well, at minimum we are going to have another day for the 1.5 respin. 17:21:36 so everone knows, the criterion is "The boot menu for all supported installer and live images should include an entry which causes both installation and the installed system to use a generic, highly compatible video driver (such as 'vesa'). This mechanism should work correctly, launching the installer or desktop and attempting to use the generic driver." 17:21:36 what if we kick this to final and no-one can figure out a fix by then? 17:21:39 adamw: So, if we punt this, are we ready for the release, do you think we have gotten enough testing? 17:21:47 would you actually want to block final on it or would you then be in favour of 'eh let's just drop the criterion'? 17:22:16 mboddu: if we agree to delay decision to tomorrow to do some smoke testing on 1.5, i think we could go 17:22:19 i would block final on this, i think. unless someone makes a convincing argument that the set of users affected is sufficiently small 17:22:30 I'd be in favor of killing this criterion. 17:22:37 I'd kill the criterion 17:22:45 I think we need the criteon if we have the option... 17:22:48 The set of graphics adapters it covers is an ever-decreasing set. 17:22:57 what i'm getting at is, if you wouldn't really block final on this, we shouldn't just punt the criterion to final, we should remove it. 17:23:01 because people will select it and tell us we suck when it doesn't work 17:23:15 * mboddu agrees with nirik 17:23:20 i guess also note that this bug is, i think, gnome-specific? 17:23:23 nirik: I think we should also look at dropping this as a selectable (read, not manually entered) item on the bootloader 17:23:31 so in theory the option might actually work on kde 17:23:35 (and non-blocking desktops) 17:23:57 lovely. :) 17:24:26 "Its not working on our default desktop, try our spins" :) 17:24:35 mboddu++ 17:25:09 this is a pretty complex one... lots of axies. ;( 17:25:15 well, i mean, that's a consideration for removing the option 17:25:26 so for the purposes of today, should we provisionally drop the criterion for beta and let QA make the final decision as part of their regular business? 17:25:35 bcotton++ 17:25:41 i.e. QA can decide if it's a beta, final, or dropped criterion 17:26:00 well, I think what we end up doing may also depend on how hard it is to fix? 17:26:10 well, it's not solely up to QA to decide, but aside from that, that's reasonable 17:26:44 I think, we either fix it now (which I am inclined to) or remove the criteria. Since it doesn't sound good that our default desktop has issues while others work 17:26:57 agree that we at least believe this should no longer be a beta criterion and leave further decisions for later 17:27:09 proposed #agreed We will consider the criterion for the issue provisionally dropped for beta and make the final decision as part of regular business 17:27:10 mboddu: unfortunately we can't fix it right now. :P 17:27:17 at least not unless you have ajax in a cage over there...=) 17:27:17 yeah, let's postpone it to final criterion and decide later 17:27:22 adamw: I have faith in you :P 17:27:56 I guess I'm ok with removing it from beta and redeciding it, but we should set a deadline for that for before final freeze.. 17:28:03 so we aren't discussing it in the final go/no-go 17:28:22 nirik: agreed. what's a reasonable deadline? 17:28:23 * bowlofeggs didn't realize we had the authority to cage people… 17:28:27 muahahaha 17:28:30 huh, i may have to revisit my evaluation that f29 had the exact same bug 17:28:41 i don't think it did any more. oh well 17:28:44 it clearly had *some* bug... 17:28:56 bcotton: week before final freeze? to give time to propose critera change and get it added/removed? 17:29:38 that would be 9 April just for the record 17:29:49 modified proposed #agreed We will consider the criterion for the issue provisionally dropped for beta and make the final decision as part of regular business by 9 April 2019 17:29:55 personally, I think we need to either remove the option and critera, or fix the option and keep the critera... but others might disagree. 17:30:18 I agree :) 17:30:19 i am okay with having non-working option in beta 17:30:27 bcotton: +1 17:30:35 bcotton, ack 17:30:45 ack 17:31:21 nirik: I agree with that. 17:31:25 nirik: I agree with that 17:31:33 whoa 17:31:38 sgallagh: haha :) 17:31:40 okay, i don't see any objections, so 17:31:44 #agreed We will consider the criterion for the issue provisionally dropped for beta and make the final decision as part of regular business by 9 April 2019 17:31:51 and that's the end of the blockers 17:32:07 anything else re: blockers before we move on? 17:32:46 sgallagh: ^ do you have anything related to cockpit issue? 17:33:05 mboddu: No issue exists. 17:33:06 Or are you still digging into it 17:33:14 sgallagh: Okay 17:33:22 nvm then 17:33:23 No, it turned out to be a networking issue. Once that was resolved, Cockpit was fine 17:33:30 nirik: uh, wait, did you mean you don't want to ship beta with the option present but broken? 17:33:43 sgallagh: Ahh okay 17:33:58 #topic Current status — Release candidate compose 17:34:14 bcotton: I am about kick of 1.5 compose 17:34:28 #info Beta-1.4 is the current Release Candidate 17:34:37 adamw: no, I meant final 17:34:46 #info Beta-1.5 is about to begin to address only the outstanding blocker 17:34:49 ok 17:35:29 mboddu: when would you expect the compose to complete? 17:36:01 bcotton: I think like 9 EDT 17:36:18 #info Beta-1.5 should complete this evening EDT 17:36:55 adamw: how much additional testing would need to be done? the change in question is pretty unimpactful, i expect 17:36:56 bcotton: It took 7.5 hrs for 1.4 17:37:56 just openqa tests and some quick media smoke tests should be enough 17:38:08 awesome 17:38:19 anything else on the compose before we move to test status? 17:38:26 oh great 17:38:36 so this nomodeset bug is also, i think, *fedora-specific* 17:38:48 la la la, i can't hear you 17:39:20 Since the important people are around, can I get +1 FE on https://bugzilla.redhat.com/show_bug.cgi?id=1691257 if possible 17:39:27 Its a easy fix and we will get 1 more image 17:40:02 +1 FE 17:40:07 #topic 17:40:09 #topic (1691257) fedora-arm-mate f30 spin failed to build 17:40:09 Its not going to affect anything, just increasing the disk size 17:40:10 #link https://bugzilla.redhat.com/show_bug.cgi?id=1691257 17:40:12 #info Proposed Freeze Exceptions, spin-kickstarts, NEW 17:40:17 +1 FE 17:40:21 +1 FE 17:40:22 there's another one too. 17:40:32 if you are entertaining FEs 17:40:48 nirik: 1691189? 17:40:57 yeah 17:41:08 +1 FE for fedora-arm-mate 17:41:24 I'm fine with 1691257 since it can't affect the rest of the compose or change the results. 17:41:38 +1 FE for fedora-arm-mate 17:41:38 nirik: Thanks guys, well, I need to make the PR for f30, since they made it for master (dont want to repeat the same mistake) :D 17:41:54 #agreed BZ 1691257 AcceptedFreezeException 17:42:03 sure 17:42:41 #topic (1691189) Please build Design Suite for Fedora 30 Beta 17:42:43 #link https://bugzilla.redhat.com/show_bug.cgi?id=1691189 17:42:45 #info Proposed Freeze Exceptions, spin-kickstarts, NEW 17:42:59 * nirik mailed the spin list a list of failures yesterday, glad to see some folks investigated and proposed FE's/fixes 17:43:05 +1 this one as well. 17:43:06 +1 FE 17:43:19 +1 FE 17:43:22 +1 FE 17:44:05 #agreed BZ 1691189 AcceptedFreezeException 17:44:20 okay, test time 17:44:32 #topic Current status — Test matrices 17:44:39 #link https://fedoraproject.org/wiki/Category:Fedora_30_Test_Results 17:44:55 #undo 17:44:55 Removing item from minutes: 17:44:58 #link https://fedoraproject.org/wiki/Category:Fedora_30_Beta_Test_Results 17:46:01 i'm seeing a lot of green here 17:46:07 * nirik is testing Xfce now 17:47:02 i think we're basically covered except a couple of ARM tests (if pwhalen didn't do them yet) and SAS, as per ye olde traditions 17:47:59 any other questions or comments on testing? 17:48:09 I think I have the upgrade tests left, will double check 17:49:30 pwhalen: did you want us to hold for your check or were you providing an update? 17:49:50 bcotton, no need to hold, I get them done 17:49:58 rock 17:50:04 so now we get to the exciting part 17:50:20 #topic Go/No-Go decision 17:50:43 so clearly we don't have an RC that can go at the moment. we have three choices 17:50:46 well, technically we have a blocker still... should we keep the meeting open and revisit once we have the new compose? 17:50:47 * bowlofeggs looks around to see if the go/nogo cops are watching 17:50:58 1. say I'm sure Beta-1.5 will be good. we're go 17:51:12 2. declare no-go and reconvene next week as scheduled 17:51:24 3. declare a pad hold and reconvene tomorrow 17:51:43 I am inclining to the third option 17:51:51 3rd one 17:51:55 agree 17:52:04 3 17:52:13 bowlofeggs: Haha :D 17:52:22 option 3 falls outside of our normal procedure, but we can make reasonable deviations. holding a week because of the current blocker seems excessive 17:52:37 👮 say 3 17:52:39 Option 3 also has precedent 17:52:40 3rd option 17:52:42 yeah, option 3 is ok for me 17:52:45 I'm good with 3 17:52:52 no it isnt we have done this before 17:52:54 so the only argument i'd have against 3 (and i see it's the popular choice) is: do we have enough slack in the tasks that happen between now and 10am EDT Tuesday? 17:53:01 2++ (which is 3 in C) 17:53:12 * mboddu is kicking off the Beta 1.5 now 17:53:12 bowlofeggs: No it isn't. ++2 would be 3 in C 17:53:14 (meaning, i also like 3 and am trying to be too clever) 17:53:28 sgallagh: haha right 17:53:41 okay, 3 has overwhelming support it seems 17:53:46 bcotton: provided we can reuse testing from current compose I think we might be ok 17:53:53 it mostly falls on QA folks... 17:54:19 proposed #agreed We will delay the decision until 1700 UTC on Friday 22 March 17:54:29 * bcotton checks to see if this channel is available then 17:54:35 yeah, i'm not expecting us to retest everything 17:54:42 * mboddu checks the lorax BRO 17:54:45 I'll do the xfce tests on arm again, apologies for the error. 17:54:47 since it's almost identical to 1.4 just doing basic checks should be enough 17:55:00 yeah, I guess we'll have to test mainly if images are booting/are installable? 17:55:20 * nirik just tested x86_64 Xfce, looks fine. 17:55:26 nirik++ 17:55:26 frantisekz: Karma for kevin changed to 38 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:55:32 is 1700 UTC enough time for QA to do the testing, assuming most folks will be asleep when 1.5 completes tonight 17:55:45 yeah, I think we're good 17:55:48 i'm thinking particularly of NA West Coast folks like adamw 17:55:59 * sumantro concurs with frantisekz 17:56:11 * mboddu will test kde 17:56:32 i don't do most of the testing these days :P 17:56:47 the previous compose finished around the same time, european tz folks got the testing done 17:56:50 #info adamw isn't important anymore 17:56:56 this is correct! 17:57:00 i only did hw raid test 17:57:14 yeah, we don't have hw raid hardware here... 17:57:28 okay, then i'll poll the responsible parties to formalize the decision 17:57:52 proposed #agreed We will delay the decision until 1700 UTC on Friday 22 March. We will meet in #fedora-meeting-1 17:57:56 FESCo? 17:58:15 Concur 17:58:17 ack 17:58:30 RelEng? 17:58:56 * nirik changes hats 17:58:57 ack 17:59:03 QA? 17:59:08 ack 17:59:11 ack 17:59:14 #agreed We will delay the decision until 1700 UTC on Friday 22 March. We will meet in #fedora-meeting-1 17:59:29 Thanks nirik 17:59:31 Fedora 30 Beta is Fedora 19 17:59:48 mboddu: figured you were firing the compose off. ;) 17:59:58 nirik: Yup, kicked it off 17:59:59 #action bcotton to announce delayed decision 18:00:09 #action bcotton to schedule next Go/No-Go meeting 18:00:14 #topic Open floor 18:00:23 Anything else we need to cover? 18:00:33 #info Release Readiness meeting is at 1900 UTC in #fedora-meeting-1 18:01:46 thanks to all 18:02:17 okay, i'll see you all in an hour back here. and in 23 hours back here again :-) 18:02:38 #endmeeting