15:00:17 #startmeeting Fedora QA meeting 15:00:17 Meeting started Mon Sep 28 15:00:17 2015 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:17 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:22 #meetingname fedora-qa 15:00:22 The meeting name has been set to 'fedora-qa' 15:00:26 #topic Roll call 15:00:33 ahoyhoy folks, who's around for some meetingy fun? 15:00:42 I am. 15:01:07 morning bruno 15:01:09 how's it going? 15:01:33 * tflink is here 15:02:00 morning tim 15:02:31 I'd like to warn that doing a dnf distro-sync --allowerasing --excludes='kernel*' does do what one would expect. You end up with no kernel installed and if you reboot before fixing it you get to have lots of fun. 15:02:48 * satellit listening 15:03:22 I'm going to file a bug for this. I'd have done it already but was having some issues with the new bugzilla. 15:03:44 brunowolff: hum, interesting that the package protection thing doesn't kick in there 15:04:40 * roshi is here 15:05:06 hi roshi 15:05:13 we're short the czech folks today as it's a holiday there 15:05:42 o/ adamw 15:06:07 Nope. I was doing an upgrade from f22 to f23 and the kernel protection was blocking the upgrade. So I figured I'd just exclude the kernel. It didn't work out well. The affected machine wouldn't boot from dvds or flash drives and my existing Fedora cds were too old to work with luks encryption. 15:06:31 fun! 15:07:03 brunowolff: fix it in dracut rescue mode? 15:07:08 oh, no, that needs an initramfs... 15:07:15 #topic Previous meeting follow-up 15:07:48 sigh, time for me to be embarrassed again 15:08:00 #info "adamw to find slide deck from Flock 'exploratory testing' talk and post it to the list" - yeah, still didn't do that. let's retire it this week. if i ever remember, i'll still do it... 15:08:24 #info "adamw to set up Cinnamon Test Day asap" - pants, didn't do that either, it somehow shuffled off of my todo list. i'll do it this week. been told that cinnamon is working quite well, though 15:08:26 I eventually found a working CD burner and I had one CD-R blank left and put the netinstll image on it and fixed it. 15:08:33 #action adamw to set up Cinnamon Test Day 15:08:51 #info "adamw to get new anaconda build from sbueno and get RC1 (or failing that TC6) built soon" - yep, that happened, Beta is all done. 15:09:10 #info "kparal to ensure DNF upgrade issues are properly considered by relevant developers / committees / whatever" - there's been ongoing discussion on that stuff and there's a fesco ticket, we'll hit it later 15:09:15 I brought it up because it might be relevant to any unofficial instructions for using dnf to upgrade to f23. 15:09:25 "tflink to finish up getting easy-karma fixed up for Bodhi 2" - there's an update in testing, right tim?\ 15:09:46 AFAIK, yes 15:10:40 "tflink to finish up getting easy-karma fixed up for Bodhi 2" - there's an update in testing, please karma it. 15:11:25 https://bodhi.fedoraproject.org/updates/FEDORA-2015-16439 15:11:28 https://bodhi.fedoraproject.org/updates/FEDORA-2015-16438 15:11:31 https://bodhi.fedoraproject.org/updates/FEDORA-2015-16437 15:11:35 (22, 21, 23 respectively) 15:11:47 OK! any other follow-up? 15:13:11 alrighty 15:13:13 #topic Fedora 23 status 15:14:09 well, Beta is out and seems not to be causing any explosions 15:14:18 I'm planning to request a Final TC1 today 15:14:24 #info adamw planning Final TC1 request today 15:14:38 sweet 15:15:38 the blocker list is looking a bit full, 2 proposed and 12 accepted... 15:15:46 oh, there's a QA meeting happening. Thanks for the reminder in #fedora-qa, btw.... 15:15:48 some of those should drop out when a few packages go stable 15:15:51 danofsatx: sorry! 15:16:06 * adamw only wakes up 15 mins before we start and spends most of that getting coffee 15:16:55 sounds about right 15:16:57 * danofsatx points adamw to www.keurig.com/ 15:17:03 danofsatx: i have a nespresso 15:17:16 danofsatx: but dude, i need at least five minutes to remember how to press buttons 15:17:18 even better. but it doesn't take 15 minutes. 15:17:37 muscle memory. I'm just sayin'. 15:17:44 so to go over the major f23 controversies... 15:17:51 we've got a topic coming up for the dnf upgrade stuff 15:18:18 we're finally getting a workaround for the systemd-kills-everything-on-shutdown bug - the update in updates-testing ATM reverts the patch 15:18:34 don't think there's anything else major going on? did I miss anything? 15:18:52 no groups in apper? 15:19:23 * adamw hasn't looked at that one since f22 15:19:34 is it the same thing that was going on then? 15:19:51 plasma apper does updates and single files but no groups 15:20:02 may I ask about bug 1262617? 15:20:05 satellit: apper is dead, for all intents and purposes. muon is the app-stream reader for Plasma 15:20:23 will we include it/ 15:20:27 ? 15:20:30 ...and tell people about it? 15:20:37 beadle: sure, let's do one thing at a time 15:20:49 adamw: thanks 15:21:00 that's a very good question. I'll bring it up at the KDE meeting this week. 15:21:05 danofsatx: thanks a lot 15:21:23 what time/channel is that, so satellit can join if he likes? 15:22:46 here, tomorrow, same time as this one 15:22:53 k thanks 15:23:37 ok, so, beadle's bug 15:24:37 beadle: this looks like a virtualbox bug. 15:24:41 Thanks. not sure what to say except it's not a usable VM. installing vbox additions hangs. I've worked around it by kiling processes 15:24:45 it's the one providing the systemd service that's causing problems. 15:25:58 yep, definitely a vbox bug, not a fedora bug. 15:26:00 if someone can comment on the bug report to help me understand then I won't take any more of you folks' time. but I dont understand that comment 15:26:27 beadle: the way you initially wrote the command was wrong 15:26:36 ? 15:26:49 beadle: so to the systemd devs it looked like you were complaining about 'systemctl vbox-add', which would be an internal systemd command and thus a bug for them 15:27:02 oh 15:27:04 but there *is* no 'systemctl vbox-add' command, which is why they closed the bug as an erroneous report 15:27:35 the command you actually ran was 'systemctl start vboxadd-service' or 'systemctl start vboxadd', which is just telling systemd to start a particular service 15:27:55 the service itself is provided by the VirtualBox packages, and the bug is likely in there, not in systemd. so it looks like a valid problem, but not a bug in systemd 15:28:36 only thing is, virtualbox is not provided by Fedora. so i'd recommend you report this problem to wherever you got virtualbox from, or possibly ask about it on the mailing list (as there are other virtualbox users there who may be able to help) 15:28:40 * satellit I have vbox 5.04 hee on f23 (Oracle f22 rpm and additions and it works 15:28:45 here* 15:30:43 alrighty...any other f23 stuff before we move along? 15:30:49 I did not know that the call was to something provided by virtual box. I did think you folks had more influence but it seems I'm the only one seeing this problem? that is a huge surprise. I thank you all for your help. 15:31:07 #action danofsatx to bring up status of KDE graphical packaging tools (apper vs. muon) at KDE meeting 2015-09-29 15:31:23 beadle: we have no particular influence over virtualbox, i'm afraid 15:31:28 sorry we can't help more 15:32:40 adamw: I needed to understand and this helped so very much. Thanks to you and to all! 15:32:51 #topic Upgrade package handling 15:33:01 so, this seems like one of the big outstanding issues for F23 15:33:25 the situation is that dnf-plugin-system-upgrade basically works, but its default handling of some package dependency issues is unfortunate 15:34:16 it silently ignores available updates that cannot be applied for dependency reasons, so you can get an apparently-OK upgrade transaction that actually misses out many packages and leaves your system half-upgraded, which we know from past experience is a bad idea 15:34:40 in theory all packages should have correct deps so the mixed-up resulting system should still run, in practice...it's not always the case 15:35:07 there is a blocker bug and a fesco ticket for this: https://bugzilla.redhat.com/show_bug.cgi?id=1263677 , https://fedorahosted.org/fesco/ticket/1481 15:35:14 so it should come up for discussion at the next fesco meeting 15:35:26 i just wanted an agenda topic in case we had any further info or suggestions for fesco 15:35:59 distro-sync --allowerasing seems to do a better job, except for the kernel. 15:36:41 just that minor little problem :P 15:36:58 i'll adjust the wiki page text to mention --allowerasing as an option 15:37:10 i still like advising use of --best on the first run, to see what the problems are... 15:37:32 +1 15:38:06 I have not tested with multiple kernels does it work correctly? 15:38:13 I filed an RFE months ago that failure to remove a running kernel should not block the entire transaction. I have a new bug today for the --exclude='kernel*' issue. 15:41:36 ok, sounds like we don't have anything further to send to fesco, but i'd encourage anyone who's interested to show up to the fesco meeting on wednesday 15:41:41 i'm figuring this will be up for discussion there 15:42:00 upgrade --best doesn't do downgrades. And we aren't great about making sure things are always upgrades. 15:49:53 alrighty, so, we're all going monday drinknig? 15:50:11 what do you mean "going"? I thought we already started 15:51:31 i was just trying to preserve some kind of illusion 15:51:50 the cake is a lie. 15:51:56 the whiskey is not 15:51:59 alright, thanks for coming along folks 15:52:51 please do tell me if i'm missing stuff out or whatever 15:53:04 or, you know, flaming torches and pitchforks work fine 15:53:06 #endmeeting