16:04:24 #startmeeting F22-blocker-review 16:04:24 Meeting started Mon Mar 30 16:04:24 2015 UTC. The chair is roshi. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:04:24 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:04:24 #meetingname F22-blocker-review 16:04:24 The meeting name has been set to 'f22-blocker-review' 16:04:24 #topic Roll Call 16:04:29 who's around? 16:05:02 * pschindl is here 16:05:18 ahoy 16:06:13 nazdar! 16:06:19 #chair pschindl jreznik adamw 16:06:19 Current chairs: adamw jreznik pschindl roshi 16:06:23 * oddshocks here 16:06:35 * oddshocks was staring at the meeting rooms again :P 16:07:32 hehe 16:07:59 #topic Introduction 16:07:59 Why are we here? 16:07:59 #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. 16:08:03 #info We'll be following the process outlined at: 16:08:05 #link https://fedoraproject.org/wiki/QA:SOP_Blocker_Bug_Meeting 16:08:08 #info The bugs up for review today are available at: 16:08:10 #link http://qa.fedoraproject.org/blockerbugs/current 16:08:13 #info The criteria for release blocking bugs can be found at: 16:08:16 #link https://fedoraproject.org/wiki/Fedora_22_Alpha_Release_Criteria 16:08:18 #link https://fedoraproject.org/wiki/Fedora_22_Beta_Release_Criteria 16:08:22 #link https://fedoraproject.org/wiki/Fedora_22_Final_Release_Criteria 16:08:25 hey, the list grew since last night... 16:08:27 7 beta proposals 16:08:32 first up: 16:08:32 #topic (1206760) KDE desktop doesn't notify for available updates 16:08:33 #link https://bugzilla.redhat.com/show_bug.cgi?id=1206760 16:08:33 #info Proposed Blocker, apper, VERIFIED 16:09:51 +1 from me 16:09:55 Sooo that plasma-pk-updates update will fix it? 16:09:58 pretty clear and has a fix already 16:10:06 yeah 16:10:17 +1 anyhow 16:10:20 +1 16:10:21 +1 16:10:31 +1 16:11:06 proposed #agreed - 1206760 - AcceptedBlocker Beta - This bug is a clear violation of the beta criterion: "Release-blocking desktops must notify the user of available updates, but must not do so when running as a live image." 16:11:19 ack 16:11:27 who's secretarying? i'll do it if needed 16:11:38 danofsatx, you around? 16:11:44 ack 16:11:48 ack 16:11:54 #agreed - 1206760 - AcceptedBlocker Beta - This bug is a clear violation of the beta criterion: "Release-blocking desktops must notify the user of available updates, but must not do so when running as a live image." 16:12:10 looks like it's you then adamw - unless we have other volunteers 16:12:33 alrighty 16:13:03 #topic (1206420) docker missing from 22 Beta TC5 x86_64 trees 16:13:03 #link https://bugzilla.redhat.com/show_bug.cgi?id=1206420 16:13:04 #info Proposed Blocker, distribution, ON_QA 16:14:12 this caused the server DVD fails in tc5, and non-dep-complete DVDs are blockers by policy anyhow. 16:14:41 +1 16:14:43 seems clear 16:14:48 +1 16:15:22 proposed #agreed - 1206420 - AcceptedBlocker Beta - This bug is a clear violation of the following criterion: "When installing with a release-blocking dedicated installer image, the installer must be able to install the default package set." 16:15:26 ack 16:15:32 * oddshocks nods 16:15:45 ack 16:15:52 * kparal lurks, ping me when you get to my/lbrabec's bugs 16:15:52 we'll have no foreign nodding practices at this meeting, damnit, oddshocks 16:15:57 you'll +1 and ack like God intended 16:16:06 :P 16:16:07 ack 16:16:29 heh 16:16:46 #agreed - 1206420 - AcceptedBlocker Beta - This bug is a clear violation of the following criterion: "When installing with a release-blocking dedicated installer image, the installer must be able to install the default package set." 16:16:56 #topic (1207251) Fedup fail to decrypt disk 16:16:57 #link https://bugzilla.redhat.com/show_bug.cgi?id=1207251 16:16:57 #info Proposed Blocker, fedup, NEW 16:16:58 aye and nay! 16:17:22 well,which is it? :P 16:17:41 it was to the previous nodacking discussion :D 16:17:56 +1, i think (though it'd be good to confirm) 16:18:03 +1 16:18:13 we can close it if it is ok 16:18:18 +1 16:18:19 +1 16:18:35 pschindl: with it, I'm +1 16:18:36 adamw: :P 16:18:48 proposed #agreed - 1207251 - AcceptedBlocker Beta - This bug is a clear violation of the following beta criterion: "For each one of the release-blocking package sets, it must be possible to successfully complete an upgrade from a fully updated installation of the previous stable Fedora release with that package set installed" 16:19:17 ack 16:19:28 ack 16:19:55 #agreed - 1207251 - AcceptedBlocker Beta - This bug is a clear violation of the following beta criterion: "For each one of the release-blocking package sets, it must be possible to successfully complete an upgrade from a fully updated installation of the previous stable Fedora release with that package set installed" 16:20:09 #topic (1206404) Crash on transition from g-i-s to GNOME on 'basic graphics' install (nomodeset) 16:20:12 #link https://bugzilla.redhat.com/show_bug.cgi?id=1206404 16:20:14 #info Proposed Blocker, gdm, NEW 16:20:21 this one's a bit borderline, but i thought i'd put it up for discussion. 16:21:17 Not sure if I should vote on this, on account of never having heard of g-i-s 16:21:19 :P 16:21:28 * roshi reads 16:21:29 oh 16:21:31 initial setup 16:21:42 yeah :) 16:21:42 there's second g-i-s crash, I'm not sure if the same or not but sounds similar in terms of description - no user, no luck 16:21:42 * oddshocks 9:21 am here ;) 16:22:00 pretty much all the packages you'll see in a blocker meeting are user facing installation packages 16:22:11 oddshocks: it's the initial setup thingy that runs when you install with GNOME and don't create a user. 16:22:18 you also see a version of it after first log in to GNOME with a new user. 16:22:25 in this bug you get to see both! 16:22:44 Always love having a good selection of bug cases to enjoy 16:23:06 if it at least creates the user I would be more for FE 16:23:34 yeah, same here 16:23:39 and if it happens only in nomodeset then I'm +0.8 FE/+0.2 blocker 16:23:44 it does create teh suer 16:23:55 maybe even 0.9/0.1 16:24:08 "On reboot, gdm runs and you can login to GNOME" 16:24:52 +1 FE 16:25:24 0 blocker, +1 FE 16:25:27 on reboot, gdm runs - check. And you can login to gnome - yes, but you have to reboot again :) (common bugs if it survives till release?) 16:25:27 only nomodeset 16:27:13 ehm. If you are thinking about what I just wrote then don't do that it was meant for another universe where it makes sense :) 16:27:40 I'll +1 for FE 16:28:52 pschindl: oooh, how do I get there? 16:28:54 proposed #agreed - 1206404 - RejectedBlocker AcceptedFreezeException Beta - This bug doesn't quite violate the criterion (gdm/g-i-s runs, and a user gets created) and appears to be only reproducible with a specific setup. If this is found to be more widespread, please repropose. 16:29:12 ack 16:30:14 is that a good enough explanation? 16:30:30 sure, wfm 16:30:30 ack 16:30:34 ack 16:30:36 ack 16:30:50 #agreed - 1206404 - RejectedBlocker AcceptedFreezeException Beta - This bug doesn't quite violate the criterion (gdm/g-i-s runs, and a user gets created) and appears to be only reproducible with a specific setup. If this is found to be more widespread, please repropose. 16:31:08 #topic (1205534) gnome-initial-setup crashes upon selecting language 16:31:11 #link https://bugzilla.redhat.com/show_bug.cgi?id=1205534 16:31:13 #info Proposed Blocker, gnome-initial-setup, NEW 16:33:01 so, this is fixed or non-reproducible? 16:33:28 Seems unclear 16:33:28 I think that we need more informatin/testing 16:33:30 i haven't tried it yet 16:33:43 +1 punt for more info 16:33:49 +1 16:34:00 +1 punt 16:34:20 sure 16:35:02 proposed #agreed - 1105534 - Punt - We need more information about this bug before we can determine it's blocker status. 16:35:21 ack 16:35:37 ack 16:35:48 #agreed - 1105534 - Punt - We need more information about this bug before we can determine it's blocker status. 16:36:02 #topic (864198) grubby fatal error updating grub.cfg when /boot is btrfs 16:36:05 #link https://bugzilla.redhat.com/show_bug.cgi?id=864198 16:36:08 #info Proposed Blocker, grubby, ASSIGNED 16:36:38 lemme check if the anaconda folks have any specific thought on this 16:36:53 ah, this bug 16:38:29 * adamw waits for response in #anaconda... 16:38:41 +1 from me 16:38:48 +1 from me too 16:38:49 criteria are clear, IMO 16:39:49 +1 16:41:35 seems like ban is more preffered by dlehman 16:42:36 yeah, +1, just wanted to check 16:42:43 seems like they don't mind it being a blocker 16:44:17 proposed #agreed - 864198 - AcceptedBlocker Beta - This bug is a clear violation of the criterion: "The installed system must be able to download and install updates with the default console package manager." when /boot is on btrfs. 16:44:41 ack 16:44:54 ack 16:45:15 ack 16:45:17 #agreed - 864198 - AcceptedBlocker Beta - This bug is a clear violation of the criterion: "The installed system must be able to download and install updates with the default console package manager." when /boot is on btrfs. 16:45:28 + 16:45:36 #topic (1206394) Error: g-bd-md-error-quark: Failed to parse mdexamine data (0) 16:45:39 #link https://bugzilla.redhat.com/show_bug.cgi?id=1206394 16:45:41 #info Proposed Blocker, libblockdev, MODIFIED 16:46:50 +1 16:47:11 i'm just about to go test the fix for this 16:47:13 +1 16:48:11 +1 16:48:39 proposed #agreed - 1206394 - AcceptedBlocker Beta - This bug is a clear violation of the beta RAID criterion: "The installer must be able to detect and install to hardware or firmware RAID storage devices" 16:48:48 ack 16:49:00 ack 16:49:07 ack 16:49:38 ack 16:49:49 #agreed - 1206394 - AcceptedBlocker Beta - This bug is a clear violation of the beta RAID criterion: "The installer must be able to detect and install to hardware or firmware RAID storage devices" 16:49:56 onto the proposals for Final 16:50:14 #topic (1200901) invisible mouse cursor in wayland login-screen when in VM 16:50:16 kparal: ping 16:50:17 #link https://bugzilla.redhat.com/show_bug.cgi?id=1200901 16:50:19 #info Proposed Blocker, gdm, NEW 16:50:26 * kparal is here 16:50:30 pschindl: thanks 16:51:50 is everyone seeing this? i don't actually know that i have 16:52:05 I have seen it. 16:52:07 I see this, and lbrabec reproduced easily 16:52:15 I haven't tried it, but it sounds like a blocker to me 16:52:18 just make sure you have Tablet added to your VM config 16:52:30 Tablet device is added automatically if you create VM as Fedora 16:52:40 if you create it as Generic, the tablet device is not there 16:52:46 i have one 16:52:56 i'd have to recheck, i can't say for sure i'm not seeing the bug 16:53:10 i'm having a hard time saying we'd block release for this, esp. as you don't really hit gdm in lives 16:53:14 adamw: can you just boot it and see if you have a mouse cursor in gdm? 16:54:18 I think it should be reproducible by everyone. only if you have some arcane old VM config maybe not. but if you create a fresh new VM, you should reproduce it 16:54:18 oh, lives... 16:54:29 I just booted my VM and it is really invisible. 16:54:49 what about switching from the spice gfx? 16:55:09 is having a tablet as a default input device still valid? I don't recall doing that, usually use virt-install, and haven't seen the problems that not-tablet did in the past 16:55:17 I don't know, but I think that's not the point. if you do some adjustments, you can remove the tablet device and you're done... 16:55:42 randomuser: the tablet device allows vm fluent capture and release of mouse cursor 16:55:54 so you don't need to press Ctrl+Alt to release it 16:55:59 I wouldn't block on this, commonbugs with the workaround 16:56:02 so sounds like easy to workaround if needed, not sure this is beta blocker 16:56:12 jreznik: it's not proposed for Beta 16:56:14 oh, that's used for that functionality? 16:56:21 we're on final 16:56:22 roshi: afaik 16:56:26 huh 16:56:27 kparal, ok, nvm, virt-install did give me tablets, carry on 16:56:28 kparal: ah, sorry, I'm blind 16:56:28 TIL 16:56:45 kparal: sorry, i'm working on two other blockers atm... 16:57:37 I think we can consider this reliably reproducible 16:57:42 so far everyone has reproduced it 16:58:27 as for the impact, this is extremely annoying if you want to power off the machine from the login screen 16:58:31 it's extremely hard 16:58:47 tab? 16:58:50 but that's it, no further impact, as I see it 16:59:04 adamw: ctrl+alt+tab, if you know the shortcut. otherwise you can't switch there 16:59:08 ah, k. 16:59:21 i'm just trying to imagine the Final go/no-go meeting where we decide to slip the release for this 16:59:24 seems unlikely 16:59:38 -1, document in common bugs 16:59:45 can't you hit the virt-manager 'power' button? doesn't that trigger a clean shutdown (not 'force off')? 16:59:49 so yeah, leaning -1 16:59:53 ctrl alt to get out of a vm isn't much of a hardship 16:59:55 IMO 17:00:05 -1 17:00:05 adamw: I can try. it never worked reliably for me. sometimes it suspended the machine 17:00:25 roshi: new blocker: https://bugzilla.redhat.com/show_bug.cgi?id=1207317 17:00:31 er, new proposed blocker 17:00:53 kk, 17:00:56 I hit the power button, I see black screen, and nothing is happening 17:00:57 votes on this one? 17:01:11 VM is probably dead 17:01:23 needed to force off 17:02:35 so, I count 3 -1 17:02:45 * danofsatx finally showed up 17:02:59 I don't have a clear opinion on this. I can image we would block on it, mouse is not working. but it affects only VMs, and people using VMs can probably find a way out by themselves 17:03:01 o/ danofsatx 17:03:57 is this an F22 problem, or a libvirt/kvm/qemu problem? 17:04:05 danofsatx: wayland problem 17:04:38 proposed #agreed - 1200901 - RejectedBlocker Final - This bug isn't severe enough to violate the Final virtualization criterion and has easy workarounds. Please document workaround in common bugs. 17:04:48 .fas corey84 17:04:49 ack 17:04:49 Corey84: corey84 'Corey84' 17:04:53 (late i apolize) 17:05:00 no worries :) 17:05:28 I just got here, too, Corey84, but shhhhh!! - don't tell anyone 17:05:34 lol 17:05:48 acks nacks patches? 17:06:27 * danofsatx already acked 17:06:50 ack 17:06:51 Corey84 made too much noise coming in, though, so you probably missed i 17:06:52 ack 17:06:58 * kparal is ambivalent, that probably means ack 17:07:22 #agreed - 1200901 - RejectedBlocker Final - This bug isn't severe enough to violate the Final virtualization criterion and has easy workarounds. Please document workaround in common bugs. 17:07:26 #topic (1205649) virtualbox/nvidia driver installed makes newer kernel updates to be ignored in packagekit 17:07:29 #link https://bugzilla.redhat.com/show_bug.cgi?id=1205649 17:07:31 #info Proposed Blocker, PackageKit, NEW 17:07:36 this one is going to be much more fun 17:07:40 lol 17:07:45 hehehe 17:08:01 oh goodie 17:09:10 in short, it seems that everyone having either virtualbox, or nvidia binary drivers, or anything else from comment 12, will not get any new kernels if they use gnome offline upgrades 17:09:23 I've been seeing this on my F21 laptop for at least 2 months 17:09:31 in other words, tainted kernels, right? 17:09:43 sounds that way 17:10:00 danofsatx: it's a package manager issue, so it does not really matter whether the kernel is tainted 17:10:23 it's more related to how those package dependencies are set up 17:10:31 wait, DNF works. I'm of the opinion that DNF is default in F22, yum is no longer a blocker. 17:10:44 danofsatx: yum and dnf works. packagekit not 17:10:49 packagekit is used for offline upgrades 17:11:02 software center and it's ilk doesn't use dnf - just the same libs 17:11:03 oh, duh....this is packagekit. 17:11:07 * danofsatx needs to read more 17:11:37 the fun part is that this only happens with third-party repos, we don't have anything like that in core fedora repos 17:11:55 so many people might argue this is not a blocker 17:12:40 I would not be that sure, because virtualbox+nvidia covers a lot of people. it's true that many of them don't use offline updates, but at least some of them do. and a lack of kernel updates is an important issue, I think 17:12:54 of course, this can be theoretically fixed in the future 17:13:19 so only tainted kernels are effected then ? 17:13:20 but still, we would release F22 with a known security issue 17:13:40 Corey84: lbrabec reproduced it with fake packages which don't touch kernel at all 17:13:41 this could also break if you had your own repos set up and used a similar dep scenario? 17:13:50 roshi: yes 17:14:00 yeah, kinda a questionable one. we *have* always quite solidly stated 'we don't support 3rd party drivers', of course. 17:14:12 I said this one would be more fun :) 17:14:17 that would be kinda nuts to have happen 17:14:39 I mean, I always update with dnf/yum, so I wouldn't hit this 17:14:47 roshi: no you wouldn't 17:15:14 I'm +1 if we can find a place in the fedora ecosystem this touches 17:15:28 regardless, packagekit needs to fix this error 17:15:47 but I don't think I can say +1 blocker with the current info we have 17:16:26 -1 blocker/+1 FE, would be nice to fix it but we should be consistent with "we can't support 3rd party stuff" 17:16:40 -1/+1 FE 17:17:21 roshi: we have searched for other packaged in core fedora repos which would use this setup and we found none. because none of core packages need to inject a third-party module into kernel 17:17:26 -1 block +1 FE here too 17:17:35 ah 17:17:41 -1/+1 17:17:43 so this is not caused by tainting, but by the dep structure? 17:17:52 kernel module packages are explicitly forbidden in fedora repos iirc 17:18:06 adamw: correct, this is not about kernel tainting but about rpm deps 17:18:11 if you can't get your kernel module in the 'kernel' srpm you're not getting it in 17:18:19 so probably -1 17:19:23 proposed #agreed - 1205649 - RejectedBlocker Final - This bug only arises in 3rd party repos which Fedora doesn't support, so isn't considered a blocker. However, it would be good to fix this issue as many users could be impacted by it. Accepted as a freeze exception to get a fix in if it's ready in time. 17:19:34 also, this can be fixed with an update afaict 17:19:35 ack 17:19:59 ack 17:20:04 ackityack 17:20:10 #agreed - 1205649 - RejectedBlocker Final - This bug only arises in 3rd party repos which Fedora doesn't support, so isn't considered a blocker. However, it would be good to fix this issue as many users could be impacted by it. Accepted as a freeze exception to get a fix in if it's ready in time. 17:20:28 * kparal was again on the fence, for the record 17:20:38 onto adam's beta blocker 17:20:42 #topic (1207317) Error: g-bd-md-error-quark: malformed or invalid UUID: (null) (1) 17:20:45 #link https://bugzilla.redhat.com/show_bug.cgi?id=1207317 17:20:47 #info Proposed Blocker, anaconda, NEW 17:21:22 Wellllll....... 17:22:15 that poor blivet stikes again 17:22:27 +1 under the cited criterion 17:22:34 +1 17:22:45 Reading through Adam's BZ, every single issue I hit yesterday seems to be from the same root cause. My EFI boot partition was entered into /etc/fstab as UUID=20030-U4E211 17:22:47 +1 17:23:00 i'm about to test the *next* fix in this chain...:P 17:23:02 maybe I'll git pull again and try to help with patch 17:23:07 danofsatx: these bugs are specific to firmware RAID 17:23:31 proposed #agreed - 1207317 - AcceptedBlocker Final - This bug is a clear violation of the RAID criterion: "The installer must be able to detect and install to hardware or firmware RAID storage devices" 17:23:32 +1 17:23:36 hmmm....in that case, I have another blocker. but I need to gather more info from the system at home. 17:23:40 ack 17:23:43 +1, ack and all that 17:23:46 ack 17:23:58 #agreed - 1207317 - AcceptedBlocker Final - This bug is a clear violation of the RAID criterion: "The installer must be able to detect and install to hardware or firmware RAID storage devices" 17:24:04 #topic Open Floor 17:24:11 anyone have anything else? 17:24:17 * roshi sets the fuse 17:24:18 has anyone tried TC5 with actual DVD media? 17:24:30 not I 17:24:42 I have one point. 17:24:54 I attempted last night, and it failed to boot because it can't start livesys.service 17:24:59 https://bugzilla.redhat.com/show_bug.cgi?id=1204612 is now solved, the latest cloud build successfully goes through Tunir tests. 17:25:03 danofsatx: there's some kind of Thing with UUIDs and UEFI, IIRC. UEFI 'uuids' and linux filesystem 'uuids' are not the same thing, or something that. could be related. 17:25:14 danofsatx: the DVD is broken by the docker bug. 17:25:17 roshi, just closing the ticket is enough, correct? 17:25:36 kushal: only when the update goes stable. 17:25:50 adamw, this was just a kickstart fix 17:25:59 adamw, nothing changed in the package. 17:26:02 danofsatx, try using gdisk /dev/sdX n type ef02 pre install ( see if pre-labeling efi works 17:26:37 adamw: this wasn't a server DVD, it is KDE Live. 17:26:55 danofsatx: oh, i see. sorry, 'dvd' is confusing 17:27:19 in which case, no, haven't tested that yet, could be it takes too long and the service gives up waiting? 17:27:28 kushal, as soon as the fix is published and known working - then it can close 17:27:33 I'm going to try on a couple more pieces of hardware to confirm, it could be the stupid samsung laptop. 17:27:35 kushal: in that case, yeah, if it's tested then it's OK to close the bug 17:27:39 danofsatx, when you bz up ping me think I may have an idea 17:27:45 adamw, roshi thanks :) 17:28:33 3... 17:28:44 2.... 17:29:11 1... 17:29:15 thanks for coming folks! 17:29:36 wait for it .... 17:29:38 #endmeeting