17:00:40 #startmeeting F31 Final Go/No-Go meeting 17:00:40 Meeting started Thu Oct 17 17:00:40 2019 UTC. 17:00:40 This meeting is logged and archived in a public location. 17:00:40 The chair is bcotton. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:40 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:40 The meeting name has been set to 'f31_final_go/no-go_meeting' 17:00:42 #meetingname F31-Final-Go_No_Go-meeting 17:00:42 The meeting name has been set to 'f31-final-go_no_go-meeting' 17:00:47 #topic Roll call 17:00:59 * pwhalen is here 17:01:03 * pbrobinson is here 17:01:15 * satellit_ listening 17:01:26 hello, p{whalen,robinson}, satellit_ 17:01:46 also hello to pbrobinson, who was not included in the expansion above :-) 17:02:13 * pwhalen looks for probinson.. he might be the clone we ordered 17:02:19 .hello2 17:02:20 frantisekz: frantisekz 'František Zatloukal' 17:02:28 welcome frantisekz 17:02:35 Hi :) 17:02:45 .hello mohanboddu 17:02:46 mboddu: mohanboddu 'Mohan Boddu' 17:03:17 hi mboddu 17:03:33 * mboddu waves at bcotton 17:04:22 yo yo yo 17:04:26 .hello adamwill 17:04:27 adamw: adamwill 'Adam Williamson' 17:04:31 ahoj adamw 17:04:44 ahoyhoy 17:05:21 .hire more yaks for testing 17:05:21 adamw hires more yaks for testing 17:05:28 .hello2 17:05:30 .fire yaks 17:05:31 o/ just peeking in:) 17:05:32 lruzicka: lruzicka 'Lukáš Růžička' 17:05:35 adamw fires yaks 17:05:49 we have too many yaks already. if they'd stop testing, we'd stop having bugs :p 17:05:59 welcome lruzicka and spotz 17:06:12 Haha, that is true 17:06:16 thanks bcotton 17:06:20 okay, let's get this party started 17:06:24 #topic Purpose of this meeting 17:06:25 #info Purpose of this meeting is to check whether or not F31 Final is ready for shipment, according to the release criteria. 17:06:31 bcotton, but you would not be able to drink milk and burn droppings :) 17:06:40 #topic Purpose of this meeting 17:06:41 #info Purpose of this meeting is to check whether or not F31 Final is ready for shipment, according to the release criteria. 17:06:45 no, bad copypasta 17:06:46 #undo 17:06:46 Removing item from minutes: INFO by bcotton at 17:06:41 : Purpose of this meeting is to check whether or not F31 Final is ready for shipment, according to the release criteria. 17:06:47 * mboddu sees party and grabs a beer :P 17:06:48 #undo 17:06:48 Removing item from minutes: 17:07:06 #info This is determined in a few ways: 17:07:08 #info 1. No remaining blocker bugs 17:07:09 #info 2. Release candidate compose is available 17:07:11 #info 3. Test matrices for Final are fully completed 17:07:12 yay 17:07:34 1. No, 2. No, 3. No 17:07:36 #topic Current Status — Blocker bugs 17:07:43 mboddu: right? should be an easy decision 17:08:00 #info 5 Proposed Blockers 17:08:02 #info 5 Accepted Blockers 17:08:14 +1 mboddu i'm going for a drink 17:08:25 :D 17:08:30 bcotton: #info nogo, #endmeeting and done :D 17:08:36 so i'm inclined to go ahead and do a blocker review for the 5 proposed blockers so that we can get the label slapped on them before monday. objections? 17:08:43 fiiiiiiiiiiiiiiiine 17:08:47 i'll bring my drink here 17:08:53 mboddu++ 17:09:03 * pbrobinson gets a G&T 17:09:06 cheers adamw 17:09:06 can't we just ship latest nightly ... :) ? openqa tested it... so 17:09:15 #topic (1761777) multilib issue with bashbug 17:09:16 #link https://bugzilla.redhat.com/show_bug.cgi?id=1761777 17:09:18 #info Proposed Blocker, bash, NEW 17:09:47 -1 Blocker, +1 FE - which I mentioned in the bug 17:09:54 yeah, i haven't seen a justification for blocker here 17:10:38 yeah, -1 blocker, is there a fix proposed already? I am a bit hesitant with FE... 17:10:50 they seem to be arguing about the fix 17:11:10 there's a ticket to update the multilib blacklist 17:11:10 https://pagure.io/releng/issue/8906 17:11:11 can it be shipped as an update post release? 17:11:13 I'd either punt or -1 FE before I see what's agreed to be a fix here 17:11:34 or does it affect live images and related 17:11:47 I think adding it to multilib should fix it 17:12:31 which, i think requires an FE by releng policy 17:12:32 I merged multilib PR in rawhide, but unfortunately we couldn't get a rawhide since then due to other reasons 17:12:39 so I'm +1 FE for a blacklist update 17:12:53 they say that it is better to fix, because it might be worse fixing later, so I am +1FE 17:13:06 i think the blacklist is applied at compose time 17:13:15 Right 17:13:16 yeah, that sounds fine 17:13:18 so if we apply the change before final compose, bash-devel will be left out of the x86_64 repo for the compose 17:13:22 so +1 FE 17:13:25 adamw, so we want it for RC 17:13:26 if we don't, it will always be there in the frozen release repo 17:13:27 yeah 17:13:35 +1FE 17:13:36 +1 FE 17:13:37 er, bash-devel.i686 i mean of course :) 17:13:39 -1 blocker, +1 FE 17:14:54 prooposed #agreed 1761777 - RejectedBlocker(Final), AcceptedFreezeException(Final) - This does not appear to violate any release criterion, however it is accepted as a Freeze Exception to avoid having an uninstallable package in the compose. 17:15:05 ack 17:15:07 ack 17:15:19 ack 17:15:37 ack 17:15:43 ack 17:15:49 ack 17:15:54 #agreed 1761777 - RejectedBlocker(Final), AcceptedFreezeException(Final) - This does not appear to violate any release criterion, however it is accepted as a Freeze Exception to avoid having an uninstallable package in the compose. 17:16:14 #topic (1691430) dnf.exceptions.Error: Incorrect or unknown "arch": armv7hcnl 17:16:15 #link https://bugzilla.redhat.com/show_bug.cgi?id=1691430 17:16:17 #info Proposed Blocker, dnf, POST 17:16:58 This is +1 blocker, and fixes are available, but not sure if dnf reviews/approves in the time frame 17:17:06 as the proposer this is critical for support for ARMv7 on a bunch of HW/virt environment inc next gen builder HW 17:17:13 * nirik is here now. 17:17:13 And I am not sure how invasive the fixes are 17:17:24 I've talked with dnf guys today, they-re not sure about the safety of the fixes 17:17:25 the fixes aren't very invasive 17:17:43 Oh thats good to know 17:17:47 but they said they-re willing to merge them downstream if I remember correctly 17:17:53 frantisekz: Whats their reasons? 17:18:04 i'm kinda willing to be +1 on this if pbrobinson and nirik say so (i don't entirely understand the bug) but would prefer minimal change 17:18:20 I didn't talk too much about the details, just quick talk in the office around the cubicle 17:18:21 i still don't entirely understand the details of the bug 17:18:22 frantisekz: yes, but the other distro has revealed the reasons for their change and IMO it's no sustainable 17:18:31 iiuc, this just goes back to the way it was before the patches were added 17:18:42 I've had some internal threads about some of it 17:19:09 basically they made the stack check for some features on aarch64 hardware. 17:19:17 it doesn't revert it 100% but it fixes things to stop the problems we're having and the issues about supportable on arm platforms 17:19:29 I've also had people from Arm itself verify they are correct 17:19:36 +1 blocker 17:20:20 I've dropped a couple of the patches that were incorrect architecturally but doesn't directly affect and it's old platforms we don't care about 17:20:37 I don't know how widespread the affected hardware is... but I know for sure it affects the lenovo ampere stuff, which is enterprise gear... one of the only ones. 17:21:05 yeah, I am willing to vote for +1 17:21:12 seems like there's a general consensus on blocker here, what criterion do we want to base it on? 17:21:21 if pwhalen who usually does all the testing says +1, I throw my +1 in 17:21:47 bcotton: there's a number, virt, installing, updating. 17:21:55 I'm a bit sad that dnf is so different in f31. ;( But oh well. 17:22:23 this is because DNF team strive to make one DNF to rule them all 17:22:37 does this only affect VMs? that's one thing i was not clear on 17:23:20 well, but it's not... f30 and f32 are one version and f31 is an older version... but anyhow, that doesn't affect this bug that affects all of them, it just makes testing/fixing it harder 17:23:25 adamw, I have only seen it on VM's 17:23:26 adamw: there's the potential there to affect things like RPi4 but we don't support that in F-31 ATM 17:23:32 adamw, I think it affects certain boards, at least this is my understanding 17:23:34 I think it's only vm's 17:23:47 because the kernel there is a complete shit show and I've not got drunk enough to deal with it 17:24:13 proposed #agreed 1691430 - AcceptedBlocker(Final) - This violates several criteria including "The release must be able host virtual guest instances of the same release." 17:24:17 but it would mean people couldn't use our images with a custom kernel 17:24:26 ack 17:24:27 Just to toss in here, I'm a little concerned about 32-bit containers on aarch64 kernels too 17:24:27 ack 17:24:28 ack 17:24:36 ack 17:24:37 sure, whatever, ack :P 17:24:43 ack 17:24:55 jlinton: agreed, I need to test that too 17:24:57 good point jlinton 17:25:33 #agreed 1691430 - AcceptedBlocker(Final) - This violates several criteria including "The release must be able host virtual guest instances of the same release." 17:25:58 #topic (1762689) [abrt] gnome-software: gs_flatpak_get_installation(): gnome-software killed by SIGSEGV 17:25:59 #link https://bugzilla.redhat.com/show_bug.cgi?id=1762689 17:26:01 #info Proposed Blocker, gnome-software, NEW 17:26:06 this one doesn't seem to have much info yet 17:26:17 Yeah, we need more info on this bug 17:26:26 so far my instinct is -1, it doesn't seem like a general issue that hits the criteria, but i'd be fine with punting to monday as we clearly are gonna slip 17:26:38 +1 punt 17:26:56 any objections to leaving this for monday? 17:27:04 yeah. +1 to get more info and leave 17:27:17 I didn't see this in the wild, +1 punt 17:27:25 #agreed 1762689 - Deferred to Monday's blocker review meeting 17:27:34 +1 punt/ack 17:27:40 #topic (1760307) CVE-2019-16746 kernel: buffer-overflow in net/wireless/nl80211.c [fedora-all] 17:27:40 ack 17:27:41 #link https://bugzilla.redhat.com/show_bug.cgi?id=1760307 17:27:43 #info Proposed Blocker, kernel, ON_QA 17:27:57 +1 punt 17:28:04 soo, there is moderate on cve details, +1 FE? 17:28:21 lruzicka, was that to the previous bug? 17:28:35 yeah, the one with flatpaks 17:28:44 I think, I am +1 FE 17:28:52 -1 Blocker 17:29:13 there are enough votes to accept this one as a freeze exception, and as it's ON_QA blocker status may be moot (hopefully) 17:29:37 anyone want to state with enthusiasm that we should accept/reject it as a blocker? 17:29:45 why do not we want to block on that? 17:29:53 on what basis lruzicka? 17:30:03 I don't mind the state, I feel we should have it in the release how ever you wish to tag it :) 17:30:07 the cve criterion is only for higher impact imo 17:30:26 The reason in comment 5? 17:30:46 frantisekz: well there's no documentation as to the impact, debian has it as high because remote 17:30:50 the 'moderate' tag is somewhat suspicious to me 17:30:51 It seems plausible to me that we want it in the release. 17:30:52 +1 FE, not sure I would vote blocker without more info 17:30:56 IE it's affected by network 17:30:58 yeah, local vs. remote seems like it's not agreed on 17:31:01 RH has it as local 17:31:06 but mitre has it as remote 17:31:15 so 🤔🤔 17:31:21 i think we can accept it as FE, get the update in and forget about it, though 17:31:25 whatever... +1 FE would get it to the release anyway and that's what we can agree upon it looks 17:31:32 well it's the 802.11 stack and it's around SSID so a rouge SSID could potentially affect it? 17:31:57 but +1 FE or +1 blocker I don't mind 17:32:11 but +1 FE does not guarantee that it is in the release if there is no fix ready. Can we make sure the fix is there? 17:32:28 the fix is ready there kernel is in updates-testing 17:32:32 I agree with adamw, accept it as FFE and push the update and leave it at that 17:32:34 okay, well it's already accepted as an FE, so let's defer to Monday to see if the fix worked and if not re-evaluate blocker? 17:32:35 it may make things tricky if we have more kernel fixes and don't take this. ;) 17:32:42 ok, so we can +1FE 17:32:45 yeah, punt on blocker 17:32:59 great 17:33:00 +1punt 17:33:09 +1 fe, +1 punt on blocker 17:33:19 #agreed 1760307 - Previously accepted as a Freeze Exception. Blocker decision deferred to Monday's blocker review meeting 17:33:43 #topic (1762291) Discover does not refresh after manipulating packages which results in error messages. 17:33:44 #link https://bugzilla.redhat.com/show_bug.cgi?id=1762291 17:33:46 #info Proposed Blocker, plasma-discover, NEW 17:35:03 -1 Blocker, it seems it only happens if you remove the package really soon after installation if I understand it properly 17:35:23 -1 blocker, +1 FE... 17:35:27 kparal is now amending that to 'it may happen any time within the same run of the app' 17:35:39 but still, installing and removing in one app session seems pretty uncommon to me 17:35:47 -1 blocker , +1 FE 17:35:54 adamw, you can change your mind 17:35:58 the only case i can see for 'install and remove' is "whoops fat fingered it" 17:36:04 adamw, or hit a wrong Install button 17:36:04 which, i mean, i guess it happens 17:36:06 -1 Blocker, +1 FE as its not breaking things, just a ugly behaviour 17:36:07 yeah 17:36:15 but the fact that it's not new is an argument against blocker 17:36:19 so, i think -1 blocker / +1 fe 17:36:39 yes +1FE -1 blocker 17:37:03 as you wish 17:37:13 but don't blame me 17:37:40 you have your right to say "I told you" lruzicka :) 17:37:49 proposed #agreed 1762291 - RejectedBlocker(Final), AcceptedFreezeException(Final) - This seems to be a rare case in real usage, but it's worth having a fix in the compose. We are forbidden from blaming lruzicka. 17:37:52 such bugs could eat the whole desktop bread 17:37:53 ack 17:37:59 ack 17:37:59 ha. ack 17:37:59 ack 17:38:09 ack 17:38:21 :) 17:38:36 ack 17:38:41 #agreed 1762291 - RejectedBlocker(Final), AcceptedFreezeException(Final) - This seems to be a rare case in real usage, but it's worth having a fix in the compose. We are forbidden from blaming lruzicka. 17:39:28 okay, that's the end of the proposed blockers. i do want to mention this accepted blocker for the record 17:39:31 #topic (1747408) Cannot upgrade to Fedora 31: package exa-0.9.0-2.module_f31+5365+04413d87.x86_64 requires libgit2.so.28()(64bit), but none of the providers can be installed 17:39:33 #link https://bugzilla.redhat.com/show_bug.cgi?id=1747408 17:39:34 #info Accepted Blocker, distribution, NEW 17:39:43 such a fun one. ;) 17:39:45 :) 17:39:50 this is starting to feel like a bug that will have us slipping until thanksgiving 17:39:57 heh, now get the drinks 17:39:58 * nirik thinks we should just do the compat package and move on 17:40:08 so i am open to suggestions on how we can get everyone to agree on a solution 17:40:20 * mboddu thinks beer isn't good enough and needs a much stronger drink 17:41:19 the DNF team approached me with their solution on Monday -> they wanted to pull an error message string, but FESCo voted against this. 17:41:19 we don't have to get into this now, but if you have ideas later, i'd be happy to hear them. we need to come to *some* solution soon, preferably a good one 17:41:30 proposed solutions are in https://bugzilla.redhat.com/show_bug.cgi?id=1747408#c72 , if you don't want to scroll throughout that page 17:41:32 i think there are sufficient ideas in the bug 17:41:48 either pull the compat package into f31 or just write the workaround into dnf-plugin-system-upgrade and gnome-software 17:41:49 yeah, now we just need to figure how to to get people to agree to one of the ideas 17:41:54 and they said the would love to fix it but they needed "clear requirements" 17:42:05 i think i like the 'do the workaround in dnf and g-s' idea best 17:42:07 I think the compat package is the best answer. ;( 17:42:11 FITE ME 17:42:23 imo, I'd vote for solution #1 , seems safest not to cause any more delays 17:42:24 what exactly is a compat package? 17:42:24 👊 17:42:37 okay, well i'll continue to use my charm to try to get this worked out :-) 17:43:11 executive decision: we'll skip voting on the two proposed 17:43:11 also FESCO can create blocker for F32 to ensure we don't need to have another ugly solution for F32 17:43:26 two proposed FEs, that is 17:43:33 so i guess let's move on to the compose 17:43:34 I prefer #1 just because its a known ugly solution 17:43:45 #topic Current status — Candidate compose 17:43:49 the compat package is adding a libgit2_0.28 package to f31 (or 0.27? would have to look) 17:44:00 no candidate yet, i believe? 17:44:08 bcotton: Yes 17:44:09 nirik, an unmodular package? 17:44:17 correct 17:44:20 lruzicka: Yes 17:44:30 #info No candidate compose has been requested 17:44:32 no candidate, yeah, since we still have outstanding blockers 17:44:40 are we still composing well in general? 17:44:44 yeah 17:44:46 yep. 17:44:51 Yup 17:44:55 well at least there's that :-) 17:45:05 there are still some spins/labs that are failing 17:45:05 nightlies are composing and passing everything in openqa except the arm test which has been busted forever and a modularity check which finds a ton of known bugs 17:45:18 #info nightlies are composing and passing everything in openqa except the arm test which has been busted forever and a modularity check which finds a ton of known bugs 17:45:30 also armv7 containers have been failing since forever. 17:45:31 nirik: do the spins/labs owners know this? 17:45:34 yes, modularity usually fails on incorrect module definitions 17:45:46 I mailed out a status a few weeks ago to spins and devel... 17:46:03 nirik++ 17:46:05 May be we should send another email 17:46:09 which was great, because one of the broken ones made me see we were building f31 with rawhide kickstarts. ;) 17:46:25 yeah, wouldn't hurt to do another round before we get a candidate 17:46:27 we can, but can we get changes in now? or they would need FE's? 17:46:38 there are like 18 modules, that do not have default streams (some of the on purpose) and some of them do not have the yaml file in fedora-module-defaults. 17:46:38 which spins 17:46:41 yeah, but hey, we can do FEs. 17:47:14 https://pagure.io/releng/failed-composes/issue/331 17:47:25 #link https://pagure.io/releng/failed-composes/issue/331 17:47:40 okay, let's move on to test coverage 17:47:41 They would need FE's and we can always say if they are invasive 17:47:43 Games, Scientific-KDE, Jam-KDE, Scientific 17:47:53 #topic Current status — test coverage 17:47:53 and arm containers/server 17:48:22 as always, https://www.happyassassin.net/testcase_stats/31/ 17:48:44 #link https://www.happyassassin.net/testcase_stats/31/ 17:48:49 cloud tests haven't been run for nearly a month 17:48:52 non of those are actual blocking are they 17:48:59 none 17:49:02 yes they are 17:49:07 cloud images are still technically release blocking 17:49:29 adamw, I can do local cloud tests tomorrow 17:49:39 until we decide coreos is gonna replace them or whatever, but that hasn't happened yet 17:49:40 Southern_Gentlem: Nope, they are not blocking 17:49:40 thanks 17:49:46 yes they are :P 17:49:51 :D 17:49:52 i meant the games, scientific-kde or the jam-kde 17:50:02 ohhh sorry 17:50:04 cross purposes 17:50:05 no, those spins aren't blocking 17:50:18 ^^ which is what I thought Southern_Gentlem is asking for 17:50:19 not blocking, but it would be nice to have them for once... 17:50:29 desktop tests for ARM haven't been run since beta 17:50:38 why haven't the cloud tests been run? 17:50:51 bcotton: i mean, because no-one did them? 17:51:31 AD tests haven't been done yet 17:51:32 that would certainly explain it! 17:51:42 sgallagh, do we have anyone doing those atm? 17:52:19 adamw, I've been testing the desktop, I'll update the wiki 17:53:01 thanks pwhalen 17:53:07 pwhalen++ 17:53:07 frantisekz: Karma for pwhalen changed to 3 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 17:53:25 default boot and install tests are showing empty since beta, which suggests to me i've got a wiki reporter bug for openqa, i'll look into that...the physical media tests we'll do when we have a candidate compose 17:53:44 no-one has formally run the xen test the whole cycle 17:53:59 raid tests need doing 17:54:06 but mostly, we're looking decent 17:54:46 #info mostly, we're looking decent 17:54:50 adamw: Sorry, I haven’t done them yet. I didn’t prioritize it because it was clear we were not going this week and I have a lot else on my plate 17:55:03 sure, as long as we get them done for the candidate should be fine 17:55:13 though you know, if it's broken it'd be good to find out :P 17:55:16 I’m at a school event for my daughter right now. I’ll get on those later. 17:55:28 * sgallagh leaves 17:55:29 okay, so let's get to the heart of the matter 17:55:31 #topic Go/No-Go decision 17:55:38 not much of a decision to make 17:55:45 any objections to a "No-Go"? 17:55:48 nope 17:56:06 #agreed Fedora 31 final is No-Go 17:56:14 #action bcotton to announce decision 17:56:18 #topic Next meeting 17:56:31 So normally we do the same time the following Thursday 17:56:35 fixed the wiki reporting bug... 17:56:36 but i have a bit of a conflict 17:56:41 adamw++ 17:56:43 obvs no-go 17:56:59 so how do we feel about 1400 UTC on Thursday 24 October? 17:57:01 yeah, no go 17:57:13 * nirik is going to be out next week, but everyone else have fun. :) 17:57:24 bcotton, sounds good 17:57:36 nirik: Define fun? ;) 17:57:53 mboddu, adamw: 1400 UTC next Thursday work? 17:58:00 ugh 17:58:01 that sounds early 17:58:09 that's uh 17:58:11 That works for me 17:58:11 7am? 17:58:14 well i might be a bit late 17:58:15 oh, I've forgot about adamw :D 17:58:19 adamw: yeah, unfortunately 17:58:25 i'll do my best 17:58:28 but at least it means our Brno friends can do home earlier 17:58:41 is that 16:00 CET? 17:58:42 yeah, we'd be pretty happy about that timing :D 17:58:45 yeah 17:59:01 ok, no objection 17:59:03 let's make sure the meeting is *really long* 17:59:21 adamw, we can discuss the libgit bug 17:59:26 i'll bring some extra coffee for you, adamw 17:59:34 adamw: I got you covered for it :) 17:59:41 lruzicka: Exactly :) 18:00:01 #agreed The next Go/No-Go meeting will be *1400* UTC on Thursday 24 October 18:00:25 #action bcotton to check meeting room availability 18:00:44 #info The F31 Final Release Readiness meeting will be in 60 minutes in #fedora-meeting-1 18:00:49 #topic Open floor 18:01:03 anything else before we all resume banging our heads on our desks? 18:02:12 Nada 18:02:21 bcotton: you stopped? 18:02:38 not here 18:02:39 * Southern_Gentlem loads more bullets for the dead horses 18:02:45 adamw: just long enough to type 18:02:49 Southern_Gentlem++ 18:02:49 Southern_Gentlem, get a decent whip 18:03:03 when a problem comes along, you must whip it 18:03:08 Southern_Gentlem, maybe you can beat them to achieve more 18:03:13 okay, sounds like we've been as productive as we know how 18:03:18 thanks everyone, see you next week! 18:03:21 #endmeeting