16:00:03 #startmeeting 16:00:11 #meetingtopic Fedora QA Meeting 16:00:22 #topic locating bodies 16:00:29 mine's here 16:00:30 * poelcat here 16:00:35 Greetings folks! 16:00:43 this is our hot new meeting time 16:00:51 I suspect we'll have some confusion this week 16:00:56 heh :) 16:01:06 I'm always confused ... so that's nothing new 16:01:30 * kparal here 16:01:40 Howdy kparal 16:01:54 this isn't the AA meeting? 16:02:09 adamw: oh oops, wrong agenda :) 16:02:10 howdy guys :) 16:02:14 jeff_hann: greetings! 16:02:24 * Viking-Ice clocks in.. 16:02:31 * dpravec is here too 16:02:37 Viking-Ice: dpravec greetings gents 16:03:22 * f13 16:03:29 f13: welcome 16:03:46 do we have a wwoods lurking? 16:04:06 * skvidal is lurking 16:04:23 skvidal: welcome you lurker! 16:04:23 ;) 16:04:37 okay ... let's get started 16:04:43 #topic Follow-up from last week 16:05:06 for folks using the play-at-home version, the proposed agenda was sent to the list ... https://www.redhat.com/archives/fedora-test-list/2009-July/msg00529.html 16:05:28 Viking-Ice: thanks for knocking off the first time 16:05:29 # [Viking-Ice] - updating meeting channel page w/ new QA time 16:05:42 kparal: thanks for the assist with updating the QA/Meetings links too 16:05:46 no problem 16:06:26 okay, flying through the list ... 16:06:26 # [jlaska] - discuss ideas with fedora-infrastructure for getting test candidates out to testers 16:06:53 spoke to mmcgrath last week about some of the problems experienced in the past on the alt.fp.org system 16:07:02 he isn't aware of any new problems since a recent system upgrade 16:07:23 and has set things up to monitor for bottlenecks once the test compose #1 has been posted 16:07:44 More information is posted in the fedora-infrastructure ticket - https://fedorahosted.org/fedora-infrastructure/ticket/1554#comment:4 16:09:05 hopefully if there are problems, we'll hit them this week and we can then work with the infrastructure team on solutions 16:09:26 okay next up .. Viking-Ice ... I've moved this into an agenda item to give you some time to discuss any concerns 16:09:30 so what's happening exactly? 16:09:41 we're pushing test compose #1 to a publicly accessible site? 16:09:43 adamw: they've got phazers on stun and are monitoring 16:09:57 you betcha ... they want to feel the burn before we make any corrective actions 16:10:09 ok. have we decided to what extent we will announce it? 16:10:10 I thought we were facing bandwith/storage issues not performance? 16:10:45 adamw: yeah, QA is going forward on plan with allowing testers to contribute 16:10:55 bandwidth issues lead to performance 16:10:58 Viking-Ice: storage wasn't a concern as I'm told 16:11:04 jlaska: right, but do we announce to test-list only, or wider than that 16:11:08 and any performance issues were surprises to them 16:11:20 f13: i think viking means how is a 'system upgrade' going to help if the bottleneck was bandwidth 16:11:24 adamw: ah, let's stick with what we've done before for now ... keeping things equal (fedora-test-list) 16:11:27 OK. 16:11:44 adamw: nailed it. . 16:11:49 yeah 16:11:57 hopefully if it's going to fail ... let's fail now 16:12:04 and then we can work the problems 16:12:29 OK, just wanted to be clear on the plan. 16:12:36 adamw: no worries, thanks for asking 16:13:01 #topic Testing dracut in rawhide 16:13:30 Viking-Ice: it hasn't been a full week since our last meeting, but I wanted to give you some time to talk about any updates you had with Harald 16:13:42 give you a chance to walk through any topics that need QA attention 16:14:10 Well no changes in the front here I have not had time to note down and contact harald ( was out in the country the whole weekend which kinda screwed things up ) 16:14:43 Viking-Ice: ooh, I wish I had the same experience! 16:15:08 Viking-Ice: no worries, are there any related issues you wanted to talk about today ... or should we revisit offline (or next meeting)? 16:15:15 In the meantime dracut reach it's 0.7 version with additional several fixes added cmd options etc 16:16:12 Next meeting I would say.. 16:16:33 Viking-Ice: okay, I'll keep it on the list 16:17:37 okay, just a quick note on upcoming events ... 16:17:42 #topic Alpha test compose 16:18:08 assuming the F12Alpha list is cleared, we can expect a test compose built this Wednesday 16:18:08 I thought we were in good shape for this 16:18:16 except that apparently we haven't been having images for the last couple days 16:18:38 jlaska: we don't need the list clear for a test compose 16:18:47 we just need it clear of any super debilitating issues, like no images (: 16:18:55 f13: ah my mistake ... we do need those anaconda showstoppers cleared 16:19:07 * poelcat was under the (possibly mistaken impression) that we haven't been able to install rawhide for a week or more 16:19:15 poelcat: you are correct 16:19:16 no images being generated sounds different from the anaconda showstoppers 16:19:49 although the _fact_ that there's no images being generated recently means we can't be sure if the showstoppers are fixed, i guess 16:19:57 http://kojipkgs.fedoraproject.org/mash/rawhide-20090727/logs/pungify-i386.log 16:20:00 there was a short period of time where there were images 16:20:08 perhaps even working images 16:20:13 and during that time we logged a few installation blocker bugs 16:20:14 when the dbus issue got fixed. 16:20:29 looks like an error in the pungify-i386.log 16:21:02 f13: should this be bz'd or sent out for investigate to the you and the release engineering team as a trac ticket? 16:21:12 should be a bug on the blocker list 16:21:28 against component, um... distribution for now 16:21:31 alrighty ... if someone doesn't beat me to it, I'll get it logged after the meeting 16:22:23 so images and the 2 anaconda blockers are the install test blockers I'm aware of 16:22:55 i have reported 2 bugs in anaconda and still waiting for fresh images to verify the fix 16:23:13 kparal: are these issues you feel should be Alpha blockers? 16:23:23 bug#513175 and bug#513383 ? 16:23:26 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=513175 medium, low, ---, anaconda-maint-list, MODIFIED, ImportError: No module named decorator 16:23:26 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=513383 medium, low, ---, anaconda-maint-list, CLOSED RAWHIDE, Detecting devices crashes 16:23:30 thats it 16:23:37 youre fast 16:23:51 * jlaska has some handy links 16:25:05 both should be fixed, just waiting for new images to confirm it 16:25:12 okay 16:25:21 and probably it happens only on certain machines 16:25:56 well, you said you were using virt-manager, right? 16:25:56 that's all I had for the test compose ... Liam is driving this initiative, so stay tuned to the list for further details (https://www.redhat.com/archives/fedora-test-list/2009-July/msg00523.html) 16:26:03 that's a common 'machine' :) 16:26:06 yes, virt-manager 16:26:27 i'd consider installs in virt-manager being broken being close to an alpha blocker at least...but we can discuss it at the next review meeting 16:26:36 heck, I've been treating virt-manager as a reference platform 16:26:41 * jlaska agrees 16:26:45 yeah, our testing is going to focus on KVM guests 16:27:01 err, our automated testing is going to ... 16:27:09 so if that's broken we're back in the dark 16:27:31 right on 16:27:35 in that case it sounds reasonable to treat them with high priority 16:27:53 should we throw them on the f12alpha list now so we don't forget to talk about them at the meeting? 16:27:57 i had problems with virt-manager -> so now i am using only kvm :) 16:28:16 adamw: kparal: yeah definitely ... kparal won't be around for this friday's meeting too, so that's a good dbl check 16:28:21 ok 16:28:26 let's add them then 16:28:30 i'll do it 16:28:33 ok 16:28:35 thx 16:28:47 done 16:29:00 any other questions/concerns/thoughts/hicu on Liam's install testing plans coming up this week? 16:29:38 alrighty ... stay tuned to the list 16:29:43 now on to a related topic ... 16:29:49 #topic Alpha blocker bug day#3 16:30:02 nothing earth shattering here ... just a reminder about bug day #3 this friday 16:30:20 adamw: wanted to give you a chance to share and things that worked well, or not so well, about the last 2 events? 16:30:24 s/and/any/ 16:30:43 i think they went fine really 16:31:04 just the usual exhortation for anyone who sees a bug they think might be a blocker: put it on the list :) 16:31:05 * poelcat really appreciates the email recaps 16:31:09 that way you can be sure it'll be evaluated 16:31:20 I like meetbot :) 16:31:24 or else, come out to the meeting and mention it there 16:32:24 in keeping with last week, should we identify who has (or wants) the ball(s) for this weeks bug day? 16:32:34 * send announcements 16:32:37 * send meeting summary 16:32:38 16:32:47 anyone new interested in trying it out this week? 16:33:27 * poelcat will send announcement 16:33:33 :) 16:33:36 poelcat: danke 16:33:46 points to whomever can come up with some good ascii art to kick off the meeting 16:34:13 i can do the recap 16:34:25 adamw: thx! 16:34:25 * dpravec will try ascii art 16:34:35 nice, just keep it clean ;) 16:34:56 #action poelcat to send blocker bug meeting announce 16:35:03 #action adamw to summarize blocker meeting 16:35:17 #action dpravec to try his hand at ascii art for blocker meeting 16:35:23 * jlaska still playing w/ meetbot 16:35:26 is it for irc, right? 16:35:51 dpravec: you got it ... take a look at the SOP poelcat wrote ... https://fedoraproject.org/wiki/User:Poelstra/blocker_bug_meeting_sop 16:35:54 hmm thats hard to do right :) some people have wrong fonts 16:36:22 * poelcat notes sop is in draft status and needs to be finalized at some point 16:36:35 okay ... any other points to discuss on the blocker meeting coming up? 16:36:36 so feel free to use experiences learned thus far to make it better 16:36:44 * jlaska isn't exactly sure of a clean consistent way to track the test blockers 16:37:21 poelcat: it's 1000x better than the SOP I didn't write :D 16:38:05 unless any other questions, changing gears shortly ... 16:39:01 alright, wwoods + dpravec ... I know it hasn't been a long time since our meeting, but if there are any AutoQA topics you'd like to discuss with the team ... we've got some time 16:39:08 #topic AutoQA update 16:39:36 I'm writing a couple of longer things about this (a blog post and maybe some wiki docs) 16:39:53 wwoods: awesome! 16:39:58 but I've got 7/12 tests from the rawhide acceptance plan automated 16:40:25 last autotest packages have some problems, I posted something at https://fedorahosted.org/autoqa/ticket/9#comment:32 16:40:42 I'm gonna try to rig something up to send out status emails / logs 16:41:17 Hum cant we automate this https://fedoraproject.org/wiki/Features/XZRpmPayloads 16:41:18 wwoods: I think there support for sending mails built into autotest 16:41:26 as in the test 16:41:37 Viking-Ice: probably, yes, but that's outside the scope of israwhidebroken.com 16:41:48 wwoods: please ping me when your docs will be online 16:41:49 yup 16:41:54 hopefully the wiki docs will give a good idea about how to write a test though 16:41:58 dpravec: will do 16:42:23 jlaska: oh excellent. I'd like to have some useful public test results between now and when we get a nice web frontend for this stuff 16:42:49 I'll give that a look 16:43:20 wwoods: I could be wrong, there is some email support in the global_config.ini packaged by autotest, we might just be able to update that 16:43:38 yeah, but if all it does is (e.g.) send out links to internal URLs for logs 16:43:43 then that's not gonna be real useful 16:43:55 so I'm not yet sure it'll be that simple 16:43:56 yeah, probably not in the short term 16:44:01 xmmp reporting might be sweet 16:44:17 i mean notifications about finished tasks 16:44:32 xmpp* 16:44:33 dpravec: thanks for the update on the packaging ... looks like a file went missing from the previous autotest package 16:44:37 dpravec: yeah, eventually we want to do a big Fedora Infrastructure Message Bus thing, so we get a signal about the completion of a new compose/build/whatever 16:44:41 and tests get kicked off 16:44:49 and then signals get sent about the test results 16:44:55 and so on. we'll get there eventually. 16:45:55 f13: do you have any cycles left for an autotest packaging update, or is that something dpravec and I should look into? 16:46:06 I'll look at it today 16:46:28 f13: sweet, thanks for the turn around 16:46:41 i will test them... 16:46:52 i mean if you do new packages 16:47:11 nod 16:48:34 cool, so I think that's it for AutoQA ... are there any other comments/concerns before moving to 'open mic night' ? 16:50:07 #topic 16:50:15 I have a topic, mass rebuild 16:50:31 #topic open mic - mass rebuild 16:50:35 f13: you've got the mic 16:50:47 At some point this week, some 8K new builds will be tagged into rawhide 16:51:13 this is new arch (i686), new compression (XZ), new glibc, and new gcc. 16:51:21 should be exciting 16:51:33 we're going to need to look out for issues like this. 16:51:52 the perfect storm 16:52:04 there's a bit of a problem right now AFAIK - the rawhide RPM has been rebuilt twice 16:52:15 so if you don't already have the RPM with xz support 16:52:18 you can't upgrade 16:52:23 because all the packages are xz payload now 16:52:36 oh geez 16:52:43 didn't we go through this with the checksums too? 16:52:53 and decided that having rpm itself need new rpm was a bad idea? 16:52:53 so this will require some manual jiggery to upgrade to rawhide? 16:53:21 * f13 would strongly push for rpm itself to be built with the old compression 16:53:28 at least it's happening before the alpha 16:53:52 jlaska: well, no - the rpm package in F11 updates-testing is gzip payload with xz support 16:54:44 and using anaconda to upgrade is no problem, since anaconda will have xz support 16:54:55 the only place it's a problem is pre-rebuild rawhide -> post-rebuild rawhide 16:55:05 gotcha 16:55:33 a simple but ugly hack is to force the f11 package into the rawhide repo 16:55:47 #link https://fedoraproject.org/wiki/Features/XZRpmPayloads 16:56:36 wwoods: f13: this may have already been done ... but is this a worthwhile candidate for the rawhide blog? 16:56:50 yes 16:57:11 yeah. if you force it with --nodeps you'll hose your system (unfortunately krh seems to have found this out the hard way) 16:57:20 doh! 16:57:51 okay so http://rawhidewatch.wordpress.com ... perhaps we can add that to the Feature page "how to test"? Not sure ... how else can we warn testers 16:57:54 right now AFAICT (I have no Rawhide systems so I'm guessing) the situation will look like this: you can't upgrade anything because of depsolving problems 16:58:05 the fix is, I guess, to install the RPM package from F11 updates-testing 16:58:42 anyone interested in taking that discussion to warren? 16:59:29 that would more or less be this bug 16:59:30 https://bugzilla.redhat.com/show_bug.cgi?id=5139372 16:59:32 Bug 5139372: was not found. 16:59:34 if no takers, I'll be happy to talk with warren after the meeting and see about testing wwoods idea 16:59:43 i don't see why we can't just do f13's solution and have rpm in rawhide built with gzip compression and xz support, just like the one in f11 16:59:49 it'd only cost a few kilobytes, surely 16:59:53 skvidal: is there an extra number in that bz? 16:59:59 ? 17:00:00 rpm and any of it's deps I guess 17:00:06 hmm, true 17:00:07 https://bugzilla.redhat.com/show_bug.cgi?id=513937 17:00:09 yes the 2 17:00:09 Bug 513937: medium, low, ---, skvidal, NEW, missing dependencies 17:00:13 that's bigger =) 17:00:16 skvidal: thx 17:00:18 sorry 17:00:30 I'm not sure how hard it is to disable xz payloads on a package-by-package basis 17:00:49 well, 'all dependencies' makes it rather tricky anyway 17:00:54 rpm has lots of deps :\ 17:01:27 so stick with the f11-update 'workaround' for now? 17:01:37 adamw: enough deps that you can do a yum update rpm, yum update 17:02:53 can someone help with the summary for this? 17:03:18 everybody's screwed? 17:03:21 heh 17:03:32 that's my standard summary for everything, it's usually quite accurate =) 17:03:57 people using pre-alpha rawhide might notice that they can't upgrade due to dependency problems with rpm/glibc 17:04:08 (I don't have example messages, sorry) 17:04:22 wwoods: no worries ... and we'll recommend installing the f11-updates rpm to work around the issue for now? 17:04:42 it's because of the xz feature (link, brief explanation) - install f11 updates-testing rpm thusly as a workaround: [commandline] 17:04:54 yeah, that seems to be the recommended fix 17:05:13 okay, Ican take that 17:05:51 #action jlaska - talk to warren about a rawhidewatch post on rpm xz workaround for rawhide testers 17:06:05 okay folks ... if not other topics, let's close things out for the day 17:06:09 s/not/no/ 17:06:13 * jlaska likes extra letters 17:06:19 well maybe the livecd issue? 17:06:41 So is rawhidewatch @ wordpress supposed to be our official blog 17:06:46 ? 17:06:47 #topic open mic - livecd issue 17:07:05 Viking-Ice: official? Hrmm, I guess it's the first thing that came to mind that would get the work out 17:07:16 Viking-Ice: it's a place to report important issues with rawhide that require user intervention 17:07:23 much like this one :) 17:07:24 rawhidewatch isn't very active right now because I'm not using rawhide yet, it doesn't work on my laptop at all 17:07:28 I see so we can close the ticked in infra for official QA blog 17:07:36 no, because that's not what it is 17:07:45 qa covers more than just issues in rawhide which require user intervention 17:08:07 we wouldn't post about, oh, test days on rawhidewatch 17:08:14 yup 17:08:16 exactly 17:08:29 no then let's get a move on getting a fedora-qa wp instance from infra 17:08:51 livecd issue: the current guide on https://fedoraproject.org/wiki/QA/Test_Days/Live_Image is I believe not correct, because it uses F11 repos instead of Rawhide repos. you can see that if you follow the includes in the kickstart file. 17:08:52 were we can have rawhide status etc 17:09:05 sorry, will wait 17:10:04 Viking-Ice: hmm, maybe something to consider in the future ... it's a neat idea. But someone would need to investigate further and come back with recommendation/plan 17:10:39 kparal: i believe you have the floor 17:10:51 it's just a short note. those people building live images according to that script are in fact building images based on F11. Therefore I'm working on a best way how to modify it so that it is based on Rawhide 17:11:00 if that's what's wanted for a test day 17:11:27 kparal: indeed, thanks! You caught me cheating in that script ... I often point to a local spins-kickstart.git repo when building test images 17:11:38 :) 17:11:42 jlaska: investigate what further 17:12:00 and there are jlaska and awilliam extra repositories which i don't know if they are still needed 17:12:10 Viking-Ice: let's wait until kparal's finished, and we'll move to your topic 17:13:02 kparal: I usually toss extra packages not yet in koji on the repo url posted 17:13:15 and adamw includes rendercheck there as well for the X testing events 17:13:16 kparal: nothing in my repos is needed for anything in f12. 17:13:24 well, except rendercheck. =) i should update that build. 17:13:47 kparal: if we need to gut that page to make things clearer, feel free to give it a once over :) 17:14:06 ok. it was just a note. 17:14:22 releasing the mic :) 17:14:26 i want to thank you all for moving the meeting time from wensday (in that day i need to take care of my kids, so it was hard to pay attention to the channel) 17:14:32 heh ... releasing 17:14:42 dpravec: you're welcome, glad this works better! 17:14:50 #topic open mic - qa blogs 17:14:58 Viking-Ice: you mentioned a request to host qa blogs 17:15:22 Yes trying to track down the ticket in infera 17:15:25 in infra I mean 17:15:37 it's been there since what f9 I think with no movement 17:15:42 is this something people see a need for now? 17:16:15 Ok how are we delivering various need to know info to people? 17:16:27 by mind power ? 17:16:29 basically through every channel possible right now 17:16:38 mailing list(s), fedora planet, fedora forums etc... 17:16:57 jlaska, wwoods and I have our personal blogs on planet fedora 17:17:01 and we blog about significant qa events 17:17:05 so they get out via that method 17:17:15 Viking-Ice: you don't even blog do you? 17:17:16 ;) 17:17:21 i'm not sure whether a qa team blog would get significantly more exposure...it might, but i'm not sure how to tell 17:17:21 I dont blog 17:17:30 well microblog perhaps 17:17:33 btw, for anyone who doesn't know, anyone can get their blog on planet fedora 17:17:43 there's a wiki page with instructions 17:17:43 I know other open source projects have team blog aggregators 17:18:01 * dpravec needs to go today, see you all later 17:18:04 We should have official blog 17:18:08 dpravec: take care 17:18:33 Viking-Ice: I like the idea of organizing the blogs of testers, but I'm not ready to put lots of time+effort into a team blog just yet 17:18:35 instead of spreading the need to know info on various ( private blogs ) places on the internet 17:19:05 a qa 'planet' might be a better idea 17:19:09 is there a way to aggregate the blogs ... right 17:19:14 that are already being posted? 17:19:22 in my experience, team blogs often get neglected because no-one really takes responsibility for it 17:19:30 yeah I agree 17:19:55 but if we can continue doing what we already do ... and it's aggregated somewhere ... that seems like an easy win 17:20:09 so if we could set up an aggregator and aggregate posts from qa team members, that might work best 17:20:11 adamw: that just reflect poor management of those team 17:20:31 Viking-Ice: true, but it's hard to push water uphill - always best to look for the path of least resistance (sorry for the management cliches :>) 17:20:32 hrmm, I guess that's one way of looking at it :) 17:20:37 I think aggregation is the best way to do "team blogs" 17:20:52 the only downside of aggregation is we may get quite a lot of duplication 17:20:57 five posts about the same event for e.g. 17:21:25 well, that at least agets the word out :) 17:21:25 like we don't already get that on planet fedora? 17:21:38 Viking-Ice: does this help provide focus for your existing infrastructure ticket? 17:21:48 f13: yes, we do - that doesn't mean it's not a drawback. :) 17:22:24 team aggregation is not a solutions 17:22:34 a solution from my perspective+ 17:22:47 Viking-Ice: it's a solution that has the support of the team 17:23:58 * kparal thinks blogs aggregation can't hurt 17:24:06 Viking-Ice: do you want to update the existing ticket ... or is someone else interested in driving this with the infrastructure team? 17:24:11 then go ahead with it and close bug 489801 17:24:13 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=489801 medium, low, ---, notting, NEW, [RFE] rss reader + notification bubble 17:25:05 hmm, I'm not planning to close any bugs 17:25:16 okay, it'll be an open action item for anyone interested 17:25:38 #action - fedora QA planet aggregated blogs? 17:25:48 okay gang ... let's close it out 17:25:56 #topic closing for business 17:26:07 thanks for your time everyone :) 17:26:35 #endmeeting