16:01:18 #startmeeting 2009-11-16 Fedora QA meeting 16:01:19 Meeting started Mon Nov 16 16:01:18 2009 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:19 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:01:25 #topic gathering people 16:01:26 who's here? 16:01:32 * Oxf13 16:01:35 * jlaska is half here 16:01:47 * kparal is here 16:01:49 do we get to pick which half? 16:02:04 adamw: why not 16:02:27 we'll take the top 16:02:44 hehe 16:02:45 do we have a wwoods? 16:03:09 * Viking-Ice 50% here.. 16:03:22 you have nine and a half wwoodses 16:03:27 hey viking 16:03:34 oh wow, it's an army 16:03:37 more wwoods than you know what to do with 16:03:53 .com 16:04:09 alrighty 16:04:19 #topic previous meeting follow-up 16:04:23 so yeah, uh 16:04:28 i'm winging this 16:04:34 just a sec 16:04:47 adamw: apologies ... don't have any action items from last week ... pulling in the previous meeting tasks 16:05:02 don't see any 16:05:02 we have a few that continue to carry over. .. not sure if we want to continue that ... or drop them from the list 16:05:11 oh iswym 16:05:18 yeah we should follow up on 2009-11-02 16:05:27 * jlaska looking at https://fedoraproject.org/wiki/QA/Meetings/20091116 16:05:38 point of order, btw: shouldn't that have been wwwwwwwwwvoods? 16:06:36 Uhum Still "investigating" need to get of my lazy as and write the dam test case.. 16:06:41 adamw: ha! good point. I'll file a bug. 16:06:55 still havn't heard from milos on the ftbfs topic 16:07:02 i think we can stop carrying that one unless he comes back with it 16:07:17 sorry ... wrong channel 16:07:37 adamw: okay, sounds good, I'll drop from next weeks agenda 16:07:58 not sure what was meant by 'preupgrade test cases' 16:08:03 I've got a link ... 16:08:40 #agreed milos jacubicek's ftbfs proposal is tabled until he comes back with more 16:08:46 just wanted to follow-up from the FESCO meeting ... I've asked Hurry to take a peak at ensuring the existing preupgrade tests will capture the failure case identified this release https://fedorahosted.org/fedora-qa/ticket/30 16:09:11 the existing preupgrade test cases didn't account for the fact that *most* users running preupgrade will have 3 kernel packages installed 16:09:28 wwoods: yeah 16:09:41 ah, yeah, that. we should fix that. 16:09:59 Hurry and Kamil have some feedbackin the ticket now, I'll be replying later todya 16:10:02 today 16:10:08 while we're at it, can we think of any other things real-world preupgrade usage would be likely to hit that we wouldn't recreate from a clean install? 16:10:18 of course ... others are welcome to 16:10:19 (that we care about. so, not rpmfusion stuff.) 16:10:36 adamw: good topic ... can we move that down to the post-mortem prep? 16:10:53 I'M RUNNING THIS MEETING DAMNIT! *froths at mouth* 16:10:56 i mean, uh, yes of course :) 16:11:10 adamw: :) 16:11:33 in that case...we're done with follow-up I guess 16:11:44 did we miss anything anyone wanted to follow-up on from the last two weeks? 16:12:31 ok 16:12:38 #topic fedora 12: coming soon 16:12:51 tomorrow, right? 16:12:56 so, yeah, stickster asked me to let everyone know that since fedora 12 has no bugs, we're all fired 16:13:06 :) 16:13:11 it's been a blast, everyone 16:13:17 :) 16:13:20 yeah, tomorrow 16:13:26 hah. so let's delete this https://fedoraproject.org/wiki/Common_F12_bugs page then 16:13:39 yeah, that thing's completely wrong, i dunno what idiot wrotei t 16:13:56 so yep, that's the common bugs page 16:14:05 in case anyone isn't aware of the idea by now, we put common bugs onto it 16:14:30 there's quite a lot of things that still need to be added 16:14:36 jlaska, could you make with the Magic List? 16:14:47 https://fedoraproject.org/wiki/QA/Meetings/20091116#Common_F12_Bugs 16:14:52 22 bugs on the list 16:15:01 shiny 16:15:06 I was hoping we could discuss a method to divide up the list 16:15:20 I'm happy to take the installer issues ... if people want to divide by component 16:15:37 there's too much on the list for just adamw to complete for tomorrow 16:15:45 note that X-related ones may occasionally be actually already on the page - matej has been cleaning out whiteboard sections and shortening the commonbugs links as he thinks they're too long, so we may have to come up with a new standard and refine jlaska's search 16:15:48 so was curious if folks were interesting in pitching in a bug or 2 16:16:10 well i wouldn't want to take anyone off other vital work 16:16:21 e.g. if wwoods is still fixing up preupgrade that should take priority 16:16:28 i could get through that list if necessary 16:16:36 so in the latest preupgrade update I referenced http://fedoraproject.org/wiki/PreUpgrade 16:17:16 wwoods: so that's the page we should link to from common_F12_Bugs as well? 16:17:18 so either that page needs a link to the commonbugs entry for F12 *or* the commonbugs entry for F12 needs to link to the part of that page that describes some ways to clean up /boot 16:17:23 your call 16:17:24 * jlaska nods 16:17:35 but yeah that needs to happen as soon as possible 16:17:42 i, uh, don't see anything about /boot there? 16:18:25 that's the problem. 16:18:35 wwoods: I'd like to get my feet wet on Common_F12_Bugs ... I'll be happy to take that for after the meeting and send you a draft for comments 16:18:39 oh, okay. gotcha :) 16:18:56 bug 530541 is probably the thing to examine 16:18:57 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=530541 medium, low, ---, skvidal, ON_QA, Free space check on /boot not thorough enough 16:19:03 jlaska: sounds good 16:19:31 #action jlaska will propose Common_F12_Bugs after meeting for bug#530541 16:19:56 I'll also get the RAID one while I'm at it 16:20:02 bug#533545 16:20:03 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=533545 high, low, ---, skvidal, NEW, Fedora 11 preupgrade to F12/rawhide destroys grub on raid (warning about grub on RAID not displayed) 16:21:27 so, also on the list is cleaning F12Blocker 16:21:28 jlaska: in theory that should be fixed, but the fact that /boot-RAID requires a wired network connection should be documented, yeah 16:21:38 wwoods: okay 16:22:03 only https://bugzilla.redhat.com/show_bug.cgi?id=533621 remains on the blocker list 16:22:04 Bug 533621: high, low, ---, rvykydal, ASSIGNED, Can't Boot After F12 b2 DVD Upgrade on sytem with RAID1 /boot 16:22:40 and we should drop it, as we agreed it wasn't a blocker (it's actually not about RAID /boot but about drive order changing between install and upgrade 16:22:56 so shall we go ahead and drop it from the list? 16:23:37 are we seeing a lot of +1 's on that issue? 16:23:47 not obviously 16:23:52 okay 16:24:01 there may be other bugs filed which are actually dupes of it, but then i'd expect anaconda triage to be catching them, they usually do 16:24:07 I guess it's going off the list, since we're not respinning 16:24:12 heh 16:24:23 is there something to document here? 16:25:23 sure, we can throw in a commonbugs notice based on radek's most recent comment 16:26:10 oh it's tagged already 16:26:12 let's drop it 16:26:19 k 16:26:27 dropped 16:26:49 now the only remaining bugs on the list are themselves metabugs: so do we drop 'em from the dependencies and close f12blocker? 16:27:09 hmmm ... I don't recall 16:27:21 is that one of the release actions for bugzappers 16:27:25 move to F13Blocker? 16:27:28 should they be moved to f13blocker? 16:27:31 ouch 16:27:37 wwoods wins 16:27:50 seems logical 16:28:10 or F13Alpha? 16:28:53 no, the alpha blocker list does not necessarily contain everything on the final release blocker list 16:29:59 * jlaska now back to 100% for qa meeting 16:30:16 trackers do seem to be listed under bugzappers housekeeping 16:30:17 http://fedoraproject.org/wiki/BugZappers/HouseKeeping/Trackers 16:30:25 so i'll probably punt that to tomorrow's bugzappers meeting 16:30:33 and we'll do the cleanup there 16:30:47 typically in the past we just moved everything to the blocker/target for the next release 16:31:10 but yeah, the bugzappers will discuss/decide on that 16:31:29 #info discuss plan for F12 tracking bugs in Tuesday bugzapper meeting 16:31:51 jlaska: you wanna take over now? it's your agenda 16:32:19 adamw: sure ... but I'll need prodding to move things along :) 16:32:24 #chair jlaska 16:32:24 Current chairs: adamw jlaska 16:32:25 so we don't keep kparal too late! 16:32:27 go for it 16:32:38 onto 'qa retrospective heads up' 16:33:06 Just a heads up kind of topic 16:33:22 there may be a release-wide retrospective meeting 16:33:29 kparal: I won't be late. I am attending Caribbean night tonight :) 16:33:40 kparal: oh that's right :) 16:33:59 so ... I wanted to more just get the brains moving around what worked and what didn't for F-12 16:34:07 you may recall we did this for F-11, and previous releases 16:34:19 I was planning to kick this off to the list tomorrow inviting comments 16:34:25 for discussion in next weeks meeting 16:34:52 are there any thoughts to share on a good method for collecting the feedback for discussion next week? 16:35:18 Ideally, I'd like to get 1-2 things that people thought went well, along with 1-2 things that need improvement ... from each of us 16:35:28 and of course from folks who don't regularly attend the meeting 16:35:55 I know that more issues will surface as bug reports come in from F-12 ... but I also wanted to do this soon after release so things are still fresh in your minds 16:36:15 we can sum it up on some wiki page and copy comments from mailing list there so they won't get lost 16:36:34 kparal: good tip, we did that with the F-11 test day survey ... I can try that again for this 16:37:27 okay ... well, I'd like to collect and use this feedback to build some common goals for F-13 16:37:43 so if folks have thoughts on a good approach to solicit feedback, please catch me after the meeting 16:38:17 #action jlaska to send request for retrospective feedback to fedora-test-list@ 16:38:37 okay ... it's that time again ... 16:38:53 #topic AutoQA update - wwoods 16:39:13 wwoods: you want to start things off ... and then we'll jump over to kparal after 16:39:23 haven't really had a lot of time for autoqa stuff in the past week because of preupgrade, but 16:39:42 we're trying to get the post-koji-build hook finalized and tested so that we can start running post-build tests 16:39:43 * jlaska looks at the preupgrade elephant 16:40:05 we've got the code written and a test system ready, just need some free time 16:40:14 I also talked with the rel-eng guys about how we can prevent broken deps in the repos 16:40:27 we identified the need for a post-bodhi-update hook 16:40:49 bodhi v2.0 will have it :) 16:41:04 and talked about writing a test that checks added/dropped/changed dependencies against known requirements in the public repo(s) 16:41:51 wwoods: by 'hook' do you mean 'AMQP message/QMF event'? 16:42:00 really the only forward progress last week was a couple of issues with the post-koji-build stuff that kparal found 16:42:03 because bodhi 2 will be plugin driven, but also AMQP integrated 16:42:19 lmacken: maybe later, yes, but at the moment we're using the simplest possible watcher scripts 16:42:32 so whatever exists now - polling an RPC / RSS feed / whatever 16:42:38 is what the initial design will use 16:43:37 getting the watcher working ASAP allows us to start writing the hook and tests immediately 16:43:52 and later we can replace the watcher with a nice messagebus listener 16:44:03 Moksha comes with an AMQP 16:44:07 message consumer API :) 16:44:15 that's cool, but I need to pick your brain about what bodhi provides today 16:44:26 in it's current state, for post-update hooks? 16:44:30 nothing. 16:44:32 RSS feeds 16:44:57 okay - we'll write a watcher script that polls the feed and launches tests when new items appear 16:45:06 ok, cool. 16:45:11 that's how all the other watchers currently work. 16:45:21 inelegant but JFDI-compliant 16:45:25 wwoods: you need a theme for these .... 16:45:28 bingo JFDI :) 16:46:24 anyway I *really* want to get this preupgrade stuff cleared up so that we can have some time to work on post-build tests 16:46:36 wanna be able to talk to packagers/devs about that at FUDCon 16:46:46 wwoods: are there aspects of the koji watcher that kparal can assist with ... if it relates to rpmguard or package sanity? 16:46:58 time is running short so probably it'll be a discussion about design rather than a demo of a prototype 16:47:54 kparal's suggestion about having autoqa have a way to run tests locally is probably the best path forward 16:48:17 once that's written we can make sure rpmguard works as expected *in parallel* with setting up the watcher on the autotest host 16:48:43 perfect 16:49:43 so: 1) wwoods adds --local flag to autoqa, 2+3) polish rpmguard and get watcher working, 4) test-enable rpmguard "in production" (i.e. for all new builds, sending out public notices, etc) 16:49:52 that's the basic plan, I think 16:50:08 but item 0) is: fix preupgrade for F12 16:50:24 5) fix lots of issue we ran into 16:50:30 *issues 16:50:34 kparal: 6) success 16:50:42 7) profit? 16:50:51 nope, world domination 16:50:55 aha 16:51:08 so, some updates from me? :) 16:51:16 wwoods: anything else on your end? 16:51:19 nope! kparal: go for it 16:51:23 #topic AutoQA update - kparal 16:51:35 Well, I have sent a few patches for the autoqa, as wwoods noted. You can see them in autoqa-devel list. 16:51:52 I have also posted a proposal how to make test development easier for new developers: https://fedorahosted.org/pipermail/autoqa-devel/2009-November/000018.html 16:52:03 Out of that document jlaska has drafted an AutoQA Use Cases on the wiki: https://fedoraproject.org/wiki/AutoQA_Use_Cases 16:52:09 If you have some suggestion, please contribute. 16:52:29 The autoqa-devel list is being used more and more I would say, so if anyone is interested in autoqa development, subscribe. James also proposed that our Trac ticket changes could be sent to autoqa-devel to keep people in loop, let's see what they say. 16:52:57 wasn't sure ... I could see how it might be noise for some folks 16:53:09 but if it's helpful, it's an easy change to make 16:53:21 hopefully they will tell us 16:54:08 so that's it for this week, but after wwoods merges his work back to master, be prepared for more :) 16:54:34 kparal: looking forward to it! 16:55:05 #topic AutoQA update - misc 16:55:18 I had an extra bullet here for some misc autoqa stuff 16:55:28 some of which was already mentioned, so I'll be brief 16:55:45 #info Package autotest - https://fedorahosted.org/autoqa/ticket/9 16:56:12 thanks to kparal for help testing the updated packages, I think I'm happy with where things are. They don't fully pass rpmlint at this point, but that's not something I understood as a goal for this first phase 16:56:31 I'm planning to catch up w/ Oxf13 or mmcgrath this week for guidance on how to get those updated packages into fedora-infra yum repo 16:56:45 #info Packaging autoqa - https://fedorahosted.org/autoqa/ticket/3 16:57:03 No new updates on that front ... I'll revisit this once wwoods is happy with the big changes in the private branch 16:57:18 it basically works in current git repo ... so not expecting surprises 16:57:27 #info AutoQA Use cases draft - https://fedoraproject.org/wiki/AutoQA_Use_Cases 16:57:35 This last point kparal noted above 16:57:57 sorry for stealing :) 16:58:05 not at all, thanks for bringing it up 16:58:14 I thought this might help shape the discussion for what tasks need completing for the upcoming FUDCon discussion 16:58:45 kparal's patch around running autoqa locally helped flesh this out for me. 16:58:59 in that we may not have fully outlined how we expect people to interact (or not interact) with the system 16:59:19 so I'll be trying to add more data points and the poking kparal and wwoods for guidance on what I'm missing 16:59:28 okay ... let's move to open discussion 16:59:38 #topic Open discussion - 16:59:49 quick 15 seconds and we'll have a <= 1hr meeting 16:59:50 :) 16:59:58 nada! 17:00:00 any topics not covered in the agenda today? 17:00:04 Fedora no longer runs on the number one supercomputer in the world 17:00:20 updated https://fedoraproject.org/wiki/Is_Fedora_For_Me accordingly 17:00:43 Jaguar kicked Roadrunner to second place http://www.top500.org/lists/2009/11 17:00:50 interesting 17:00:57 heh, "the number two supercomputer" 17:01:26 unless there are any other topics ... I'll close out the meeting in 60 seconds 17:01:28 Jaguar runs Cray Linux Environment encase ye are wondering.. 17:02:40 Alrighty gang ... folks can keep discussing on #fedora-qa or the list of course 17:02:51 but we'll call it a wrap for today 17:02:59 thanks for your time and adamw for kicking things off 17:03:05 #endmeeting