14:59:57 #startmeeting Fedora QA Meeting 14:59:57 Meeting started Mon Jun 21 14:59:57 2010 UTC. The chair is jlaska. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:59:57 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:00 #meetingname fedora-qa 15:00:00 The meeting name has been set to 'fedora-qa' 15:00:05 #topic Gathering critical mass 15:00:34 * jskladan steps out of the shadows 15:00:36 Hi folks, we'll get started shortly 15:00:41 jskladan: you lurker! :) 15:01:20 morning 15:01:22 * jlaska notes, I believe we are without wwoods today ... I gather he is on a train heading south right now 15:01:28 adamw: howdy 15:01:44 i would like to propose a new QA group motto 15:01:55 from dedoimedo's F13 review - http://www.dedoimedo.com/computers/fedora-13.html 15:01:57 adamw: let's hear it! 15:02:02 "Testing Fedora is like stuffing a badger down your trousers; you don't really know when it's going to bite." 15:02:08 * kparal here 15:02:08 haha :)) 15:02:13 kparal: hi there 15:02:46 #idea New QA motto - "Testing Fedora is like stuffing a badger down your trousers; you don't really know when it's going to bite." 15:02:59 should it be tested in practice first? 15:03:14 * Viking-Ice joins in by shooting the bloody badger.. 15:03:19 Viking-Ice: welcome 15:03:41 alright, I think we have enough to get started 15:04:28 Unfortunately, today is my double booked meeting day ... I've asked kparal to take over when I need to step out 15:04:33 so everyone be nice to kparal! :) 15:04:36 #chair kparal 15:04:36 Current chairs: jlaska kparal 15:04:44 #topic Previous meeting follow-up 15:04:55 #info jlaska to check-in with lmacken on the status of https://fedorahosted.org/bodhi/ticket/424 15:05:09 * adamw throws rotten fruit at kparal 15:05:16 I don't have any updates on this ticket, I wager others in rel-eng or infrastructure might know more 15:05:20 * kparal dodges 15:05:53 * jlaska pictures Kamil dodging the flying tomato ala 'the matrix' 15:06:08 I'll try find luke and see if there are any updates 15:06:14 anyone else have news on this front? 15:06:41 alright, next up ... 15:06:44 #info adamw to forward wwoods a good explanation of the nss-softokn problem scenario to ensure autoqa catches it in future 15:07:01 Adam knocked this off the list iirc 15:07:10 yup, sent to autoqa-devel 15:07:16 #link https://fedorahosted.org/pipermail/autoqa-devel/2010-June/000703.html 15:07:42 adamw: awesome thanks, hopefully that is something 'depcheck' has covered. 15:07:48 I haven't really understood whether the test for catching these issues would be hard or easy to code 15:08:11 wwoods isn't here, but I can add a discussion topic to review your nss scenario 15:08:50 well, if it should be covered by depcheck we can wait a week for wwoods 15:09:10 #action wwoods to evaluate nss-softokn dependency problem for proper 'depcheck' coverage 15:09:12 I had the impression a new test case would be needed 15:09:35 i suspect it's something that probably isn't covered yet 15:09:55 kparal: maybe, if you have any ideas feel free to drop them in response to adam's note 15:10:09 since the key issue is you need to look at whether packages will be pushed to certain repos, as well as the packages themselves 15:10:12 * jlaska hasn't yet fully absorbed the failure ... so I'm glad you have! :) 15:10:43 well I haven't, I have just a faint idea, that's why I asked :) 15:10:53 okay, so we've got this as something to check-in on next week ... hopefully we know more then 15:11:00 Next up ... 15:11:04 #info jlaska to propose a clean-up of the Critical Path Packages wiki to be a suitable home for critpath information 15:11:32 kparal and adamw, you guys noted that the current [[Critical Path Packages]] wiki page didn't have the information you wanted (or needed to link to) 15:11:32 that page received some heavy editing in the meantime 15:12:02 yeah, I tried to fill the gaps, and then got a lot of help from alaxlan and dafrito 15:12:05 #link https://fedoraproject.org/wiki/Critical_Path_Packages 15:12:31 I also filed rel-eng TRAC ticket to see if we can get a more consistent URL for the latest critpath.txt file (see http://fedorahosted.org/rel-eng/ticket/3802) 15:12:45 that would help not having to guess where the current critpath list is 15:13:14 yea, that would be great, stable URL 15:13:24 anyhow, this page came up during the FESCO meeting last week, so I suspect others are paying attention and will be helping improve the content 15:13:28 the 'background' section seems spectacularly useless (it's not practical information, just a nugget of historical interest). other than that, looks great 15:13:29 * kparal thinks he created a ticket for in some time ago at our trac 15:13:48 kparal: ooh, sorry, pass me the ticket and I'll connect the dots 15:13:57 adamw: what kind of info would you expect to see there? 15:14:30 * kparal looking 15:14:36 well, there's two ways to look at it 15:14:42 * jlaska goes to listen mode 15:14:43 if we can't think of anything useful to put there, why have the section at all? 15:14:54 jlaska: here's the ticket I filed https://fedorahosted.org/fedora-qa/ticket/60 15:15:07 kparal: thx 15:15:37 but if you want an idea of something possibly useful to go there: an explanation of why the critpath proposal came about - what problems is it trying to solve? 15:15:55 I agree with adamw 15:16:00 adamw: okay, thanks ... I'll revisit 15:16:30 #action jlaska to cleanup (or remove) the Critical Path Packages#Background section so that it provides _some_ value 15:16:34 thanks gents 15:16:52 okay ... agenda time ... 15:16:58 #topic F14 QA Recommendations 15:17:30 nothing tremendously new here ... other than to say ... 15:17:32 #info Jlaska organized Fedora_13_QA_Retrospective feedback into a list of QA recommendations for Fedora 14. 15:17:46 I've not yet heard any alarming feedback around "good lord no!" 15:18:01 my next step is to start filing fedora-qa tickets for the main tasks 15:18:20 some of the things on the list are tasks folks are always actively working 15:18:46 so I may assign some of those tickets to you 15:18:53 don't panic" 15:18:57 s/^/"/ 15:19:17 Ideally, we can start picking off tickets from the open fedora-qa F14 ticket queue 15:19:38 and if folks are looking for ways to get involved, hopefully that queue provides a list of ideas 15:19:53 Please feel free to shout if you have any concerns/ideas/complaints 15:20:13 * kparal haven't had time to study it yet, but will provide feedback soon 15:20:37 kparal: thx 15:21:29 I'm still grappling with how to see this through, but I think I'd like a page that lists all the improvements, provides tickets to follow-up and provide improved measurement of improvements etc... 15:22:17 I'd like to have better visibility into the cool things (and the sometimes painful tasks) we all do to prepare for a great release 15:22:32 I think this will help, but as always, ideas/suggestions welcome 15:22:41 okay, next up 15:22:44 #topic Proventester Update 15:23:01 I think maxamillion has a conflict with this time slot :( 15:23:05 adamw: are you able to give an update? 15:23:41 whoops, sorry, multitasking 15:23:54 sure, partially 15:24:42 on our side, we're now just working on the instructions for proven testers, which we've been thrashing out on the list. with big thanks to Aaron Faanes for his great revisions. I need to make a few suggested changes to his last draft and then we 15:24:49 we'll probably be close to done 15:25:30 the other Thing That Needs To Happen is simply for the actual update gating system to be switched on; there's not a lot of point in having a proventesters policy and team if updates aren't actually considering proventester input 15:25:37 i believe jlaska was liasing with releng about that 15:26:23 adamw: you bet, I've got that on my radar to follow-up with lmacken on 15:26:35 adamw: I'll let you know what comes out of that 15:27:03 * jlaska Aaron Faanes wiki contributions :) 15:27:32 adamw: cool, thanks for the update 15:29:06 jlaska: I'd be willing to work on editing the Critical Path Packages#Background article with you, if you want. I'll probably end up touching it anyway while I work with adamw on proventesters 15:29:28 ah dafrito == Aaron ... welcome 15:29:32 yay dafrito 15:29:35 dafrito: many thanks for the critical path wiki changes! 15:30:14 okay, next topic was proposed by Viking-Ice 15:30:33 #topic Systemd QA preparation 15:30:53 #info Viking-ice noted that Features/systemd was approved by FESCO and wanted to propose several QA ideas. 15:31:11 this was discussed briefly on IRC by Viking-Ice and adamw ... I'm not sure if what is needed to follow-up with 15:31:24 yea basically we need to start testing this as early as possible 15:31:36 one take-away for me is to start up the wiki details of F14 test days 15:32:00 Viking-Ice: would another debugging guide, much like you did for dracut, be helpful here? 15:32:01 i've mailed lennart to ask if he'd be interested in having a systemd test day 15:32:13 ooh, debugging guide sounds like a good idea. 15:32:19 #info adamw reached out to lennart for interest on a F14 test day 15:32:44 jlaska: yup I see if I cant throw together systemd wikipage with debug guide 15:32:44 * kparal is enthusiastic about systemd 15:33:16 Viking-Ice: cool, I'd love to help, but I admit I know nothing about it. I'll be glad to assist with some wiki-foo if needed 15:34:03 atleast someone needs to read over the page due to my exceptional good english hehe 15:34:06 i like the idea, i hope the implementation process isn't too full of pain. =) 15:34:23 selinux fucked as is 15:34:24 Viking-Ice: sure, do you want to send it to test@ whenever you have something you feedback on? 15:34:29 Viking-Ice: sure, if you follow the usual process (post first version / draft of page to the list for comments) i'll be happy to proof it 15:34:58 yup I'll send something to the list as soon as I have something 15:35:25 thanks! 15:35:26 cool 15:35:38 we need to make a decision when to switch to systemd by default 15:35:51 Viking-Ice: that's more of an engineering decision than a qa one 15:35:58 #action Viking-Ice will start 'how to debug systemd problems' guide and send to list for review 15:36:26 we need to check all run level cli foo service commands etc.. 15:36:39 #action jlaska to review F14 test day wiki content 15:37:24 i'm trying to think whether there's any obvious bear traps we should watch out for in relation to systemd 15:38:13 the only ones I can think of off the top of my head are related to That Other Repo, so i'll look at that in my personal time... 15:38:14 we have pre-branch test runs of F14 content in the schedule, we could use those images for discovery of pain points etc... 15:38:38 yeah, we'll probably just run into them as they crop up. i'm just trying to think of things that may be big problems we don't quite catch with our planned testing. 15:38:51 * jlaska hands baton to kparal 15:39:10 * jlaska goes into lurk mode 15:39:40 alright, systemd finished? 15:40:15 #topic AutoQA updates (initscripts, prioritization status) 15:40:23 OK, that one's on me :) 15:40:25 jskladan go! 15:41:05 so, not much happened during the last week, but we have some news :) 15:41:37 initscripts - scott fergusson finished mysql review and he boldly goes to do some more already 15:42:09 yay 15:42:17 go Scott go 15:42:18 and i'm currently 'mentoring' two more guys, who'd like to contribute, but seem kind of helpless at the moment 15:42:43 so i wrote a huuuge email explaining the lines on the 'todo list' 15:42:57 jlaska: you should have been cc'd 15:43:06 so i'll probably put it on the wiki 15:43:23 ad AutoQA - jlaska is the man of the week 15:43:40 he worked hard on gaining some machines 15:43:52 https://fedorahosted.org/pipermail/autoqa-devel/2010-June/000687.html 15:44:07 so it seems that we'll have testing instances in no time :) 15:44:16 * jskladan likes :) 15:44:17 * jlaska notes, mmcgrath and smooge get the honors for this 15:44:25 I'm just the messenger (don't shoot me) 15:44:44 and also started with the new bodhi hook testing https://fedorahosted.org/pipermail/autoqa-devel/2010-June/000688.html 15:45:03 six servers? surely you don't need all of those, send one to me ;) 15:45:11 jlaska: just so I'm aware, you're no longer blocking on us for anything except qa2 right? 15:45:20 adamw: hah, gladly! 15:46:08 mmcgrath: I have 2 outstanding items I spoke to smooge about. 1) Creating a db1 mysql db, and 2) updating ACL's so that qa0[1-6].c.fp.org are only accessible from bastion (or from autoqa01.phx) 15:46:18 mmcgrath: I need to get a ticket in for #2 15:46:47 #info jskladan wrote an explanation how to contribute with initscripts, will put it on the wiki 15:47:00 we'll probably just have you do iptables. It'd work and is easy to make alterations to when the time comes. 15:47:06 #info jlaska gathered huge number of machines available for autoqa testing 15:47:59 and I also approached Ray Chen, who wrote quite a blog about fedora qa tools and gess what - he is thrilled to join us :) he has a great insight on virtualization, so i'll write down some status report on the virtualization milestone, put it on wiki and will get the guy in the loop :) (and we could have the 'full' virt support in no time :)) 15:48:04 #info Infrastructure team comes to QA's rescue -- thanks mmcgrath and smooge! 15:48:57 #info jskladan recruited another volunteer to help up with virtualization, Ray Chen 15:49:15 great jskladan 15:49:28 which is just about it on AutoQA, i guess... I'm currently working on some frontends and stuff (stay tuned :) ) 15:49:50 so once the machines are up'n'running we'll have some fireworks :) 15:50:00 alright, thanks for the update 15:50:39 I had some vacation but aside from it I was working on forcing some tests to be executed inside virtual machines instead of on bare metal machines 15:50:52 it looks good, I will have more updates next time 15:51:16 that covers AutoQA then 15:51:36 #topic Open discussion 15:52:01 unleash your ideas now... 15:52:17 i was wondering about archiving all the TCs/RCs 15:52:34 #info Interesting test idea for anyone interested -- http://blog.famillecollet.com/post/2010/06/20/ABI-:-stability-check 15:52:50 #topic TCs/RCs archive 15:53:15 robatino: that should be discussed with infrastructure guys I guess 15:53:28 maybe the storage space would be problem 15:53:48 delta compression such as disos could be used for the install images (except for netinst) 15:54:11 if they are available quickly that would also take care of http://fedorahosted.org/rel-eng/ticket/3575 15:54:19 what would be the benefit of such archival? 15:54:39 s/archival/archiving 15:54:43 well, we obviously had a case this week where we'd like to see how the previous builds behaved 15:54:46 bugs such as https://bugzilla.redhat.com/show_bug.cgi?id=597283 that only affect a few people and can't get caught during testing 15:54:55 oh that's true 15:54:56 the mysterious x86-64 dvd boot fail on gigabyte motherboards issue 15:55:37 ok, do you think we should file a ticket at infrastructure team? 15:55:49 to hear their opinion 15:56:26 if there is enough storage space it would be beneficial to have it available at least until a new Fedora is released 15:56:32 sure 15:56:50 who's the voluteer for writing that ticket? robatino? :) 15:57:00 *volunteer 15:57:03 yes, i could do that 15:57:23 #action robatino to ask infrastructure team to archive previous RCs/TCs 15:57:43 ok, please provide url then 15:57:53 will do 15:58:13 #topic test case: ABI - stability check 15:58:18 so, we have some idea from jlaska 15:58:23 #info http://blog.famillecollet.com/post/2010/06/20/ABI-:-stability-check 15:58:31 #link http://blog.famillecollet.com/post/2010/06/20/ABI-:-stability-check 15:59:00 jlaska: do you have some time to provide more details on that? 15:59:31 kparal: I'll see what i can dig up. At first glance, it seemed like a good idea for consideration into the future rpmguard/rpmlint redesigned framework 15:59:35 again ... very far off there 15:59:37 s/there/here/ 15:59:49 so afaict it's a tool for checking whether the ABI has changed between two upstream releases 15:59:58 right on 16:00:01 to be used if we don't trust upstream's process for identifying / notifying of ABI change? 16:00:11 because usually, y'know, we get that info from upstream 16:00:24 or to be used if we (QA) don't trust the packagers? =) 16:00:51 ok, sounds good. so we should create a ticket at autoqa trac for it? 16:01:14 to be implemented as a test case when time permits 16:01:16 I can toss this into a ticket, I don't think this applies to any active milestones ... but seemed like a good idea 16:01:34 or maybe an enhancement for our rpm* tools 16:01:45 yes, it is 16:02:09 #action jlaska to create a ticket about possible enhancement/new test case for ABI stability check 16:02:09 definitely, under the framework ideas you and skvidal discussed ... hand wave ... fast fwd to when we have time to complete that ... this seems like a good candidate test 16:02:20 ok, thanks! 16:02:36 #topic Open discussion 16:02:43 more ideas! 16:04:03 * adamw has nuthun' 16:04:23 ok, let's close it, thanks everyone for attending 16:04:27 #endmeeting