15:01:07 #startmeeting Fedora QA meeting 15:01:07 Meeting started Mon Jun 29 15:01:07 2015 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:07 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01:11 #meetingname fedora-qa 15:01:11 The meeting name has been set to 'fedora-qa' 15:01:14 #topic Roll call 15:01:23 * kparal is here 15:01:25 I say, I say, is there a chap named Roll in here? 15:02:26 * Roll_ is here 15:02:38 * brunowolff is here 15:02:47 damn, I got an underscore 15:02:59 nice. 15:03:11 well, looks like that's everything...=) 15:03:41 adamw, you are a bagel 15:03:59 i was wondering why I was so firm yet tasty 15:04:10 ha! 15:04:47 soo...anyone else around? 15:04:51 * satellit listening 15:05:46 hi satellit 15:07:15 here after realizing i was in the wrong channel... 15:07:18 hi cmurf :) 15:07:56 alright, well, we've got a few people, on with the motley... 15:08:00 #chair kparal randomuser 15:08:00 Current chairs: adamw kparal randomuser 15:08:18 #topic Previous meeting follow-up 15:08:27 this topic is in severe danger of cancellation for lack of action items... 15:09:13 anything to follow-up on from last week? 15:09:31 nothing seems to have happened on the live respins topic, at least there's nothing new in the trac ticket since my previous comment 15:10:44 * roshi is here (got nerd sniped) 15:10:59 any advance in lives working? 15:11:19 the syslinux bug? 15:11:20 satellit, if I knew that, my life would be different :P 15:11:32 * satellit : -/ 15:11:59 hi roshi 15:12:03 Is h peter anvin cc’d on that syslinux bug? 15:12:09 o/ 15:12:25 Is httpd not working going to be a blocker for any of the products? I tried looking through the server roles and didn't find any clear dependency on httpd. 15:12:28 don't think he's directly cc'ed but I know he's aware of it. 15:13:06 brunowolff: probably implied in the Cockpit requirements: https://fedoraproject.org/wiki/Fedora_23_Beta_Release_Criteria#Cockpit_management_interface 15:13:25 brunowolff: oh, it's actually required to work at Alpha, https://fedoraproject.org/wiki/Fedora_23_Alpha_ 15:13:27 iirc cockpit itself does not use httpd 15:13:28 grr 15:13:33 https://fedoraproject.org/wiki/Fedora_23_Alpha_Release_Criteria#Cockpit_management_interface 15:14:37 OK. I'll mark the bug I have open as a potential blocker and copy in the criteria. Thanks. 15:15:32 huh, i thought it did 15:15:38 randomuser may be right, though 15:15:54 'cockpit apache' is one of those google fails, too. :P 15:16:10 I don't have an instance up at the moment to directly verify, though :P 15:16:11 If you chagnged it it didn't trigger an email to me. 15:16:35 do we need/want a work around for the syslinux bug? I figure copying a few isolinux binaries from f22 media to f23 media (usb) will do it. for ISO testing in a VM the qemu direct kernel boot is the way to go, or UEFI. 15:16:45 brunowolff: so if cockpit doesn't use it i can't think of anything else offhand that would make httpd a blocker for release, no 15:17:27 oops i guess we’re still on previous meeting stuff, nvm 15:17:37 cmurf: yeah, it was going to come up in a minute. 15:17:53 but we seem to have turned it into Miscellaneous Q&A anyway :) 15:17:58 so, moving on, i guess 15:18:17 #topic Fedora 23 status 15:18:25 so, here we go, can talk about syslinux all we like now! 15:18:37 I am not sure if the bug affects cockpit. It depends how it gets started. 15:18:51 for anyone who hasn't been following the plot so far, see https://bugzilla.redhat.com/show_bug.cgi?id=1234653 15:19:07 it's a socket activated python? thing, brunowolff 15:20:14 #info there's a showstopper for BIOS boot of current Rawhide images, https://bugzilla.redhat.com/show_bug.cgi?id=1234653 . the basic cause is known but requires someone with decent skillz to debug further, pjones will do it but not until July 7, we are looking for anyone else capable. 15:20:19 The issue with httpd is that it now gets sent a window size change (presumably a new systemd feature) and httpd exists when it receives the signal. 15:20:41 (let's not bet too hard on july 7; that's the RHEL deadline, but, well, if stuff's not done, ...) 15:20:49 brunowolff: hah, fun. still, i expect it should just get fixed in the normal course of things. is there a bug report? is it being neglected? 15:20:56 pjones: so, should say July 7 or later? 15:21:06 yeah, earlier is pretty unlikely. 15:21:12 You can run httpd as root from a terminal prompt and it works fine. So I wouldn't be surprised if cockpit works. 15:21:32 pjones: OK, we'll take that as read, i don't think anyone's going to be holding you to fedora QA meeting minutes... 15:21:37 okay. 15:21:38 .bug 1212224 15:21:41 brunowolff: Bug 1212224 SIGWINCH kills httpd - https://bugzilla.redhat.com/1212224 15:22:06 `dnf repoquery --requires cockpit-ws` doesn't return anything I recognize as httpd 15:22:15 cmurf: as for workarounds - well, i'm not sure releng would want to start hacking up the nightly build scripts, and it causes a bit of a problem with source availability, i guess 15:22:34 i mean testers doing it manually in the interim 15:22:35 randomuser: yeah, from just poking around in the spec it looks like it has its own webserver or something. 15:23:02 cmurf: oh, if there's a recipe to fix a built image that way, sure, throw a note in the bz, could be helpful 15:23:15 http://ur1.ca/myh4w - package requires 15:23:38 It sounds like it isn't going to be a blocker. 15:23:55 still an embarassing thing to ship 15:23:59 randomuser: note, https://bz.apache.org/bugzilla/show_bug.cgi?id=50669 15:24:11 oh, sorry, wrong persson 15:24:14 brunowolff: ^^^ 15:24:49 * randomuser reads anyway 15:25:25 brunowolff: also note from systemd.service that it's actually configured to tell systemd to use a SIGWINCH to shut apache down cleanly 15:25:25 Lovely. I'll see if I can find a way to disable this presumed systemd feature. 15:25:38 # Send SIGWINCH for graceful stop 15:25:38 KillSignal=SIGWINCH 15:26:01 so, it may be that systemd is intentionally stopping the service for some reason; the SIGWINCH message may simply be telling you, effectively, 'for some reason the service was intentionally stopped' 15:26:16 brunowolff, what, this discussion is only about running it from a terminal? 15:26:17 i can try and have a poke at it later i guess, see if i can reproduce it. 15:26:18 meh 15:26:27 randomuser: no, the opposite of that (when he runs httpd from a terminal manually it's OK) 15:26:33 ah 15:26:44 anyhoo 15:26:53 I'll stop dragging the meeting back there 15:27:07 No, I run it from a ternimal after every reboot on my rawhide machine to start it, since systemd won't. 15:27:10 so yeah, the big news on f23 status is the syslinux bug, anyone capable of debugging bootloaders is begged to help us out with that 15:28:08 #halp help needed to debug https://bugzilla.redhat.com/show_bug.cgi?id=1234653 15:28:45 any chance to do a bare metal install testing with bfo.iso (works in VirtualBox also) on a results page for the DE's in the interim? 15:28:48 where’s the latest live? is that on the releng dashboard? 15:29:00 should be cmurf 15:30:16 satellit: sure, we can nominate a nightly and throw in some stuff about working around the bug, i guess 15:30:24 i'll get right on that after this meeting 15:30:31 : ) 15:30:41 #action adamw to nominate a nightly and add instructions to the page and list about avoiding the syslinux bug 15:31:08 download speeds in BFE mountains means I should have a work around for baremetal BIOS boot with syslinux in 45 minutes to 5 hours 15:31:10 cmurf: per https://koji.fedoraproject.org/koji/tasks?owner=masher&state=closed&view=flat&method=createLiveCD&order=-id we have an 06-28 Workstation, no KDE though it looks like 15:31:23 * satellit workstation plasma and mate work from bfo.iso 15:32:16 i’ve got https://kojipkgs.fedoraproject.org/work/tasks/4623/10234623/Fedora-Live-Workstation-i686-rawhide-20150628.iso downloading now to run on an ancient Dell 15:33:43 hum - we haven't had a boot.iso since 06-23 15:33:52 i'd better check with releng why that is 15:34:02 #action adamw to check in with releng what's happened to boot.iso generation 15:34:54 anyone have anything else on F23 status? 15:35:53 OK, well, moving on, then 15:36:03 #topic Test Days 15:36:12 so, i don't think we've done anything about F23 test days, yet 15:36:15 roshi, did you have any plans? 15:36:24 is anyone else interested in test day co-ordination this cycle? 15:36:58 * kparal looks at missing pschindl ;) 15:37:27 adamw: I'm going to send out the call for test days this week 15:37:31 today, most likely 15:37:46 poking at testcloud packaging now 15:37:46 kparal: in the bar again? sigh. :P 15:37:51 roshi: OK, sounds good 15:38:06 roshi: have you checked through the F23 Change list for ones that look particularly Test Day-worthy? 15:38:47 I haven't 15:39:03 will do when I send out the call though 15:39:41 cool. 15:40:01 anyone have any ideas for things they think should get test days? 15:40:22 I think the cinnamon spin should get a test ay 15:40:36 ...mostly because it's easy to participate in, and good PR 15:40:47 true 15:40:54 * roshi uses cinnamon quite a bit 15:41:24 could be a good gateway to QA participation for a lot of folks 15:41:38 sure, and always good to check a new spin is working OK 15:41:51 * satellit cinnamon also installs fine with bfo.iso 15:42:32 #action roshi to send out the call for Test Days this week 15:42:43 I can bring the idea up to Dan but can't commit to being around during much of the actual test day at this point 15:43:15 well, file a request and we can see where it goes, i guess? 15:43:55 ack 15:45:49 okely dokely 15:45:53 that's about all i had for the agenda 15:45:56 #topic Open floor 15:45:58 any other business? 15:46:10 * roshi has nothing 15:47:27 just a quick question - do we have blocker bug meeting today? 15:47:43 it's in fedocal but I haven't seen an email invite 15:48:08 ah, the proposed blocker is gone 15:48:10 kparal: i sent out a mail suggesting we skip it as there was only one proposed blocker 15:48:12 that answers my question 15:48:16 I went to send the invite yesterday, and adam had already suggested that we vote on the one blocker in the bz 15:48:26 ok, thanks 15:49:28 #info no blocker review today as there are no proposed blockers to review 15:50:26 welp, sounds like all quiet on the QA front... 15:50:29 * adamw sets the fuse 15:50:56 * Roll_ has no questions 15:51:13 one 15:51:23 * Roll_ tips his hat 15:51:26 maybe teams should promote elections at their weekly meeting next time around? 15:51:45 yeah, that's a good idea 15:51:51 oh, sure, it'd be worth mentioning whenever one's on 15:51:53 i tend not to notice 15:51:56 I totally forgot about the fesco elections 15:52:01 * adamw has no truck with this 'democracy' nonsense :P 15:52:02 you're probably referring to pingou's blogpost about attendance 15:52:06 saw the announce, went to vote, and then it was closed 15:52:09 so I forgot 15:52:14 kparal: yes 15:52:17 roshi: yes 15:52:25 it would be a good idea to promote it more, definitely 15:52:29 i only by pure dumb blind luck remembered 15:52:41 #agreed we should promote elections during meetings - any time you know there's an election going on, if it's not mentioned in the meeting, bring it up 15:52:51 and then asked the fedoraforum.org people to make it prominent - obviously it didn’t matter much 15:52:53 good call cmurf, thanks 15:52:54 90 votes blek 15:52:58 I often notice the announcements before the election then forget to vote during the appropriate window 15:53:10 in my case, I saw the "vote is open" post before the discussions were published on fedora magazine, so I had to wait for that. it's easy to forget about it this way 15:53:17 well the announcement shouldn’t go out until you can vote. all ducks need to be in a row 15:53:34 * kparal nods 15:53:37 interviews at the same time would also be helpful 15:53:48 but at least in my universe, the news has been very entertaining the last 2 weeks so it would have been very easy to forget about fedora voting 15:54:03 pretty much 15:54:33 sounds like it'd be worth suggesting to the Grand Elections Poobah or whoever to do something with announcements - maybe a second announcement of polls opening? 15:54:41 yes 15:54:58 but also the team meetings, just schedule it as a “go vote minions!” 15:55:31 yeah, sounds good 15:55:45 Vote Minions! 15:56:03 mignonne! 15:56:08 the Minions were running? 15:56:17 aw... I woulda voted had I known 15:56:18 or we’re grilling you directly on charcol 15:56:29 minion mignon ? 15:56:35 sounds tasty 15:56:36 filet minion 15:56:37 same thing, derivation 15:56:51 they can follow or be grilled 15:56:53 haha 15:57:12 and on THAT note, ladies and gentlemen, we shall close the meeting while cmurf is hauled off to the Terrible Jokes Dungeon 15:57:18 thanks for coming, folks 15:57:30 #endmeeting