16:01:42 #startmeeting F29-blocker-review 16:01:42 Meeting started Mon May 14 16:01:42 2018 UTC. 16:01:42 This meeting is logged and archived in a public location. 16:01:42 The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:42 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:01:42 The meeting name has been set to 'f29-blocker-review' 16:01:42 #meetingname F29-blocker-review 16:01:42 #topic Roll Call 16:01:42 The meeting name has been set to 'f29-blocker-review' 16:01:49 morning folks, who's around for the first f29 blocker review meeting? 16:02:05 * pwhalen is here 16:02:19 * satellit listening 16:02:22 .hello2 16:02:23 frantisekz: frantisekz 'František Zatloukal' 16:02:29 .hello2 16:02:30 kparal: kparal 'Kamil Páral' 16:03:08 .hello2 16:03:09 sgallagh: sgallagh 'Stephen Gallagher' 16:03:46 sorry i'm a bit late, i'm all kinds of behind this morning 16:04:34 * kparal pokes lbrabec coremodule 16:05:18 how's everyone doing 16:05:36 monday-like 16:05:40 * Southern_Gentlem battling spring pollen 16:06:23 My house is falling apart around me. 16:06:36 that...that sounds like a blocker 16:06:51 everyone sitting on the beach under coconut tree thanks to OpenQA :) 16:06:57 proposed #agreed sgallagh's house stability is a definite f29 blocker 16:07:31 Well, I neglected to mention this is because contractors are ripping out rot and repairing it. 16:07:35 But thanks :) 16:08:49 i hope you're going maximum hgtv 16:09:08 adamw: I can't do that. I have a real job. Well, a job. 16:09:12 haha 16:09:54 "I walk dogs three days a week and my wife paints crotcheted hats for pernguins. Our budget is 1.6 million" 16:10:45 annnnyway. I've wasted enough of our time, sorry. 16:10:52 hahaha 16:12:24 alllrighty, let's get rolling 16:13:04 #chair sgallagh frantisekz 16:13:04 Current chairs: adamw frantisekz sgallagh 16:13:12 #topic Introduction 16:13:12 Why are we here? 16:13:12 #info Our purpose in this meeting is to review proposed blocker and nice-to-have bugs and decide whether to accept them, and to monitor the progress of fixing existing accepted blocker and nice-to-have bugs. 16:13:12 #info We'll be following the process outlined at: 16:13:14 #link https://fedoraproject.org/wiki/QA:SOP_Blocker_Bug_Meeting 16:13:14 #info The bugs up for review today are available at: 16:13:17 #link http://qa.fedoraproject.org/blockerbugs/current 16:13:18 #info The criteria for release blocking bugs can be found at: 16:13:22 #link https://fedoraproject.org/wiki/Basic_Release_Criteria 16:13:24 #link https://fedoraproject.org/wiki/Fedora_29_Beta_Release_Criteria 16:13:26 #link https://fedoraproject.org/wiki/Fedora_29_Final_Release_Criteria 16:13:28 who wants to secretarialize? 16:14:40 well, don't all volunteer at once 16:15:54 I can take that... I'll at least have some motivation to ... automatize this :D 16:16:23 #info frantisekz will secretarialize 16:16:45 * kparal is back, at the right time (after frantisekz volunteered) 16:17:21 good timing, kparal 16:17:37 we have: 16:17:38 #info 7 Proposed Blockers 16:17:42 #info 1 Accepted Blockers 16:17:45 #info 1 Proposed Freeze Exceptions 16:17:47 (all for Beta) 16:17:54 let's get started with proposed blockers! 16:18:02 #info starting with proposed Beta blockers 16:18:03 #topic (1576699) unable to do vnc installation 16:18:03 #link https://bugzilla.redhat.com/show_bug.cgi?id=1576699 16:18:03 #info Proposed Blocker, anaconda, NEW 16:19:22 +1 16:19:53 Yeah, barring information that this is unique to lnie's setup, +1 16:19:55 yeah, looks pretty straightforward. 16:20:01 +1 16:20:06 (i keep thinking we have this in openqa, but apparently we don't) 16:20:06 +1 16:20:48 +1 16:21:11 proposed #agreed 1576699 - AcceptedBlocker (Beta) - clear violation of "When using a dedicated installer image, the installer must be able to complete an installation using the text, graphical and VNC installation interfaces" 16:21:36 ack 16:21:59 ack 16:22:03 ack 16:22:14 ack 16:22:22 #agreed 1576699 - AcceptedBlocker (Beta) - clear violation of "When using a dedicated installer image, the installer must be able to complete an installation using the text, graphical and VNC installation interfaces" 16:22:34 #topic (1577405) '' object has no attribute 'FirewallEnabled' 16:22:34 #link https://bugzilla.redhat.com/show_bug.cgi?id=1577405 16:22:34 #info Proposed Blocker, anaconda, NEW 16:23:20 unless someone can demonstrate some way this affects a release-blocking path, i'm gonna have to be -1 16:23:32 nirik: is this failing the compose 16:23:33 ? 16:24:07 * kparal agrees with adamw 16:24:18 note the firewall-y tests in openQA are passing, this isn't some general 'everything that touches firewall during install dies' bug apparently 16:24:30 Given this is failing with firewall_proxy, this may be a firewalld API bug 16:24:35 hmm 16:25:17 no, but ask patrick for more details. I didn't look at this one much. 16:25:20 adamw: Does that include the server tests with enabling/disabling firewall? 16:25:30 If so, I'm good with -1 blocker here 16:25:34 sgallagh: yes. 16:25:41 and it's too early to worry about FE, IMO 16:25:54 agreed 16:26:50 A puiterwijk appears! 16:26:55 So I do :) 16:27:08 Sorry, I'd not noticed the call for blocker bug meeting until just now :( 16:27:22 puiterwijk: Opinion is generally -1 blocker on the firewall issue. Do you have anything to add beyond what's in the BZ? 16:27:55 sgallagh: from what I could find, this happens if you don't have a "firewall" statement in your kickstart 16:28:07 Oh, interesting 16:28:15 Note: that's theory 16:28:39 adamw: Is it possible that openqa doesn't hit that? 16:28:39 I have not had time to figure this out entirely, but that is the common difference between the failing vs non-failing kickstarts 16:28:51 sgallagh: probably. 16:29:12 (basically all our kickstarts have either firweall --disabled or --enable --services=, except for the container ones) 16:29:42 adamw: Do we have any specific rules about firewall in the criteria? 16:29:43 actually...no 16:29:43 I'm personally -0 blocker. I just figured I'd file it so other people can discuss it. (I *think* it's an automatic FE... But again, wanted other people to say yay/nay) 16:30:23 https://openqa.fedoraproject.org/tests/236941 uses a kickstart with no firewall line 16:30:29 that passed 16:30:35 Okay, so then away goes my theory 16:30:53 Maybe non-installed status of firewalld? (random other theory.) 16:31:01 I'm -1 blocker for this. 16:31:01 it uses http://jskladan.fedorapeople.org/kickstarts/root-user-crypted-net.ks 16:31:09 anyway, we seem fairly -1 for now 16:31:13 Anyway, -0 for blocker yeah 16:31:17 can always revote if understanding of the bug changes 16:31:26 yea, -1 here as well 16:31:29 Which also doesn't install firewalld :) 16:31:33 Yep. And I think it might be fixed soon (hopefully) :) 16:31:46 proposed #agreed 1577405 - RejectedBlocker (Beta) - as currently understood, this does not affect any release-blocking image. If our understanding of the bug changes we will revisit this 16:31:53 sgallagh: yeah. Looking at our kickstarts, we don't do anything for firewalld. 16:31:57 adamw: ack 16:32:01 yikes, late but here 16:32:10 ack 16:32:12 ack 16:32:19 sgallagh: firewalld is in @core, so it would. 16:32:38 #agreed 1577405 - RejectedBlocker (Beta) - as currently understood, this does not affect any release-blocking image. If our understanding of the bug changes we will revisit this 16:32:50 #topic (1573873) Cogl/Clutter does not support 10-bit depth colours making some Gnome applications unusable 16:32:50 #link https://bugzilla.redhat.com/show_bug.cgi?id=1573873 16:32:51 #info Proposed Blocker, cogl, NEW 16:32:51 Ah, right. And we don't do @core in container ks. 16:32:57 Actually, we do --nocore 16:33:08 So yeah, I think firewalld related 16:33:14 Anyway, sorry 16:33:49 Oh, so it is 16:34:10 I think we established that this was a blocker in F28 16:34:16 so, I asked Lukas to file this so that it's not swept under the rug and forgotten 16:34:21 And that it was resolved *there* by dropping the failing feature 16:34:40 it was not a failing feature, the apps were failing 16:34:42 and still are 16:34:59 and we disabled the feature just to be able to release 16:35:02 Well, the feature was added and the apps could n 16:35:07 n't deal with it 16:35:12 Semantics aside... 16:35:25 my assumption was that we will not want to do that again, perhaps wrong 16:35:44 i think you're maybe overplaying the 'feature' angle here a *bit* 16:35:44 but people with HDR monitors might be angry that they can't use it because of two gnome apps 16:36:06 so that's why I told Lukas to file it so that we can discuss it 16:36:08 so we're kinda playing bureaucracy games here 16:36:22 but i don't think it's reasonable to block beta on '10-bit must be enabled' 16:36:54 that's not written in the release criteria anywhere. the software has to work, is what the release criteria say. we're not really in the business of declaring that things that can be defined as 'workarounds' must go away at some given point. 16:36:58 perhaps we should ask mesa devs whether they plan to re-enable it, and if they do, block on totem and gnome maps 16:37:19 yeah, I am not sure about blocking on beta... and regarding the final, we can discuss this later 16:37:21 personally i'm only in favour of blocking when something is actually *broken*. 16:37:35 if we get back to the broken state, fine, we have a blocker bug. till then we do not. 16:37:37 my intention is that we know about the problem in advance, the developers are aware of it, and it is not dealt with in the last week as in F28 16:38:04 sure. but there's nothing to 'deal with' right now. i agree it would be good to make sure the packagers are all on the same page about this, but a blocker bug is not the most appropriate way to do that imo. 16:38:08 the point is, if we just nack it now, and mesa devs re-enable it afterwards, we will find out again late in the cycle 16:38:14 but assuming Fedora tries not to differ feature-wise much from upstream... we shouldn't keep this workaround around forever 16:38:35 Proposal: shunt this over to the Prioritized Bugs process? 16:38:47 sgallagh: it's proposed 16:39:15 adamw: I agree we're missing a different process here 16:39:15 Sorry, by "shunt" I mean "stop talking about it here unless it becomes an issue again" 16:41:23 +1 sgallagh 16:41:26 so my suggestion would be to ask mesa devs and add gnome devs in the discussion if needed. sure, it doesn't need to be a blocker bug, it can be an email thread or a pagure ticket. this is just something we can track easily 16:41:56 i just like to keep the blocker process focused. frankly, personally it's annoying if there's an "accepted blocker" on the list i can do nothing about. 16:42:15 ok 16:42:22 we're supposed to work through the list of accepted blockers getting them fixed, for this...what do we do? there's nothing to fix. 16:42:52 there's a discussion to be started, so that we figure out whether there's something that needs to be fixed 16:43:12 I don't think that's that unusual for blocker bugs 16:44:06 so, not to pick on kparal, but does anyone besides kparal think this ought to be a blocker? 16:44:13 just want to make a decision and move on here :) 16:44:16 I'm not saying that 16:45:14 well, i just mean, if no-one else wants to argue +1, we kinda have a decision 16:45:24 I figured that if mesa devs say that they will re-enable rhb10 support, then this is a blocker, yes, and if they don't (in the near future), we can drop this 16:45:38 the goal was just to prevent late surprises 16:45:43 i understand that. 16:46:00 i still don't think it's necessary. 16:46:03 alright 16:46:13 we're all cc'ed on the bug now. we're going to notice what happens whether it's an accepted blocker or not. 16:46:16 I'm not insisting, just explaining 16:46:27 right. i'm just trying to get a vote done and move on. 16:46:56 mesa devs are not CCed on that bug 16:47:12 then we should fix that :P 16:47:18 👍 16:47:23 if they read bugzilla 16:47:39 we should add PrioritzedBugs to the blockerbugs app, so we won't need to have this as a Blocker to make it more visible, ..., anyway, -1 Beta Blocker from me 16:47:51 frantisekz: good idea, patches welcome :) 16:48:13 yep... added to my todo list :) 16:49:15 proposed #agreed 1573873 - RejectedBlocker (Beta) - as things stand, nothing is broken for F29 Beta. There is no blocker bug here. If the 10-bit support is reintroduced in Mesa without affected applications being fixed, we will revisit this. 16:49:31 ack 16:49:36 ack 16:49:40 ack, mostly 16:50:23 ack 16:50:28 #agreed 1573873 - RejectedBlocker (Beta) - as things stand, nothing is broken for F29 Beta. There is no blocker bug here. If the 10-bit support is reintroduced in Mesa without affected applications being fixed, we will revisit this. 16:52:01 #topic (1575650) initial-setup fails - ImportError: cannot import name 'USER' 16:52:01 #link https://bugzilla.redhat.com/show_bug.cgi?id=1575650 16:52:01 #info Proposed Blocker, initial-setup, MODIFIED 16:52:35 +1 16:52:36 +1 blocker 16:52:37 https://openqa.fedoraproject.org/tests/236905 looks like this is fixed. 16:52:38 +1 16:53:04 +1 16:53:34 #info openQA testing indicates this is now fixed with initial-setup-0.3.61-1 , we will close the bug 16:53:49 #topic (1575626) libdnf does not handle module stream updates properly 16:53:49 #link https://bugzilla.redhat.com/show_bug.cgi?id=1575626 16:53:49 #info Proposed Blocker, libdnf, NEW 16:54:21 +1 with the criterion change referenced in comment #2 16:54:34 +1 16:55:02 +1 16:55:28 +1 16:56:11 proposed #agreed 1575626 - AcceptedBlocker (Beta) - clearly violates the newly revised criterion "The installed system must be able to download and install appropriate updates with the default tool for the relevant update type in all release-blocking desktops (e.g. default graphical package manager)" 16:56:22 ack 16:56:25 ack 16:56:45 ack 16:58:06 #agreed 1575626 - AcceptedBlocker (Beta) - clearly violates the newly revised criterion "The installed system must be able to download and install appropriate updates with the default tool for the relevant update type in all release-blocking desktops (e.g. default graphical package manager)" 16:58:16 #topic (1574711) pki-tools cannot be installed on current Rawhide (due to tomcat version bump) 16:58:16 #link https://bugzilla.redhat.com/show_bug.cgi?id=1574711 16:58:16 #info Proposed Blocker, pki-core, MODIFIED 16:58:27 this is probably fixed by now... 16:58:38 let's see how freeipa is failing in openqa atm (i'm sure it's too much to expect that it'd be *working*) 16:58:52 good lord, it works 16:58:54 it actually works 16:59:00 * adamw clutches at air, passes out 16:59:07 :D ... 17:00:59 adamw, give it a week 17:01:02 #info openQA FreeIPA tests on current Rawhide pass, indicating this is now resolved, we will close the bug 17:01:03 Southern_Gentlem: :P 17:01:11 woot! 17:01:59 #topic (1575797) Bootloader installation fails (BIOS) or installed system boot fails (UEFI) when blivet-gui partitioning used during install 17:01:59 #link https://bugzilla.redhat.com/show_bug.cgi?id=1575797 17:01:59 #info Proposed Blocker, xfsprogs, NEW 17:02:18 so, cmurf thinks this is down to XFS. i'll try and do a special run of the tests against a non-Server image today to see if they work that way. 17:02:33 still, it seems fairly clearly a blocker. 17:02:46 * adamw brb, call of nature 17:03:34 +1 17:03:38 +1 17:04:14 +1 17:05:45 +1 17:06:16 +1 (Sorry, got a phone call, back now) 17:07:46 * adamw back 17:08:18 proposed #agreed 1575797 - AcceptedBlocker (Beta) - accepted as a violation of the custom partitioning criteria, which clearly imply that a system installed with custom partitioning must actually boot 17:08:27 ack 17:08:33 ack 17:08:49 ack 17:09:18 ack 17:09:44 #agreed 1575797 - AcceptedBlocker (Beta) - accepted as a violation of the custom partitioning criteria, which clearly imply that a system installed with custom partitioning must actually boot 17:10:08 OK, that's all the proposed blockers 17:10:19 there's just one proposed FE, let's just vote on it since it's here, even though we're early 17:10:24 #info going to the one proposed FE 17:10:26 #topic (1577405) '' object has no attribute 'FirewallEnabled' 17:10:26 #link https://bugzilla.redhat.com/show_bug.cgi?id=1577405 17:10:26 #info Proposed Freeze Exceptions, anaconda, NEW 17:10:31 adamw: well, I think that this might be an automatic blocker? 17:10:36 Per the " Bugs which entirely prevent the composition of one or more of the non-release-blocking images required to be built for a currently-pending (pre-)release" 17:10:38 Did I miss the libdnf blocker? 17:10:48 sgallagh: we already accepted it. 17:10:51 Errrr, FE* 17:10:52 oh ok 17:10:58 sgallagh: if you mean 1575626. 17:11:22 puiterwijk: oh, you're right. i forgot we had those. 17:11:34 OK, I missed it 17:11:36 I just filed it as FE to get your say on that 17:11:40 #info this is in fact an automatic freeze exception per https://fedoraproject.org/wiki/QA:SOP_freeze_exception_bug_process#Automatic_freeze_exceptions 17:11:50 puiterwijk: "These bugs can be marked as accepted by any member of one of the stakeholder groups without formal review." 17:11:54 puiterwijk: so, next time you can do that. :P 17:12:04 #info we will mark as Accepted. 17:12:13 #info taking a quick look at the accepted blocker 17:12:19 #topic (1572916) kernel after 4.17.0-0.rc2.git0.1.fc29 waits for random entropy on boot 17:12:19 #link https://bugzilla.redhat.com/show_bug.cgi?id=1572916 17:12:19 #info Accepted Blocker, kernel, NEW 17:12:25 adamw: yes, that's what I thought, but as said, I wanted to ask you, given the "currently-pending (pre-)release" part read to me as "close to release" 17:12:40 so, this is an accepted blocker...anyone know what the current state of the art on this is? 17:12:43 puiterwijk: yeah, i tend to ignore that. :P 17:12:56 puiterwijk: we could probably re-word it a bit. 17:13:06 adamw: we have virtio-rng inlined which alleviates things on some VMs if you have virtio passed in 17:13:22 For hardware, it's probably just "wait long enough and it will most likely boot" 17:13:46 For VMs that 1. have dracut-fips (i.e. any of the composed images), 2. no virtio, it's still blocking boot forever until you get a cat on the keyboard 17:14:22 So... I think not much changed since last review, except maybe the virtio driver inlined? 17:15:06 right. i'm kinda wondering exactly what if anything more the kernel team is planning to do, but sounds like we don't know. 17:15:16 this one makes testing difficult. vm's are fine but arm/aarch64 on serial doesnt have the cat option 17:15:47 Yeah, I think the upstream kernel conversation went dead. The downstream libgcrypt conversation also seems to have died 17:15:51 (at least from what I'vce seen) 17:15:53 I really don't know why in 2018 every board doesn't just have a noise-generating chip on it. 17:16:19 most do. 17:16:23 sgallagh: I guess "components cost money" 17:16:25 but we still have old hardware to deal with. 17:16:45 and VMs. 17:17:04 Also, it might very well be that a lot of boards (like e.g. X-Gene) have things in separate kernel modules, which aren't loaded yet 17:17:07 VMs at least can be chained to their hypervisor's entropy 17:17:25 So it might be that we need to inline modules for more rng chips 17:19:02 sgallagh: can be, yes. the problem is they aren't by default. 17:19:07 (generally speaking.) 17:19:19 puiterwijk: i also wonder what the impact of rngd.service is exactly 17:19:30 adamw: on an X-Gene system I tried this morning: significant. 17:19:36 i mean, what works and what doesn't before rngd.service is started 17:20:16 (sidenote on VM hypervisor entropy chaining: make sure your virthost has enough entropy.... if the host has like 20 bits of entropy, guests aren't going to get much of that) 17:20:35 anyway 17:20:41 if we don't have much news, we don't have much news 17:21:03 #info the dimensions of this problem are becoming more well-understood over time, but we still don't know exactly what further (if anything) the kernel team plans to do about it at this point 17:21:11 #action adamw to ask kernel team about their plans for this 17:21:28 #topic Open floo 17:21:30 #topic Open floor 17:21:36 any other business, folks? 17:22:23 can't think of anything 17:24:13 * pwhalen has nothing else 17:25:19 alrighty, thanks a lot for coming along 17:26:14 #endmeeting