15:01:53 #startmeeting Fedora QA meeting 15:01:53 Meeting started Mon Apr 11 15:01:53 2016 UTC. The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:53 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:01:53 The meeting name has been set to 'fedora_qa_meeting' 15:01:59 #meetingname fedora-qa 15:01:59 The meeting name has been set to 'fedora-qa' 15:02:03 #topic Roll call 15:02:11 .hello dmossor 15:02:12 danofsatx: dmossor 'Dan Mossor' 15:02:13 danofsatx: just getting my cuts and pastes in order 15:02:18 * satellit listening 15:02:33 danofsatx: you probably thought i had some sensible macro or whatever for starting the meeting, right? nope. i just go to last week's meeting log and cut/paste the lines. :P 15:02:44 * danofsatx is filing unemployment claim, carry on 15:02:51 * pschindl is here 15:03:02 i believe tflink said he was sick, so we're down a flink 15:03:07 *sumantro is here 15:03:08 * kparal is here 15:03:40 sumantro__: the way you do that is to type "/me is here" :) 15:03:44 * garretraziel is here 15:04:00 #chair garretraziel danofsatx 15:04:00 Current chairs: adamw danofsatx garretraziel 15:04:01 * pwhalen is here 15:04:39 adamw :) my bad .. will keep in mind the next time 15:04:58 * cmurf lurks 15:05:18 Where are the monuts? HMM? 15:05:28 quick introduction: sumantro is one of the new interns on the RH Fedora QE team 15:05:41 cmurf: haha, there's a line-up around the block 15:05:43 * kparal waves at sumantro__ 15:05:52 thanks for making it to the meeting sumantro :) 15:05:53 Hey All !! 15:06:18 adamw, its my pleasure to be here with you all :) 15:06:36 welcome sumantro__ :) 15:07:09 thanks pwhalen :) 15:07:27 alllllllrighty then, rolling along 15:07:49 giddyup lil' pony 15:08:37 #topic Previous meeting follow-up 15:09:07 sumantro, FYI - the way we handle this usually is we go through each action item from the previous meeting and check in on the status, then the meeting leader summarizes with a #info 15:09:31 it's mostly just quick check-in at this point, if we need more detailed discussion we usually have a topic later 15:09:44 got it ! 15:10:11 so topic 1: "sumantro and arvind (new RH interns) to set up Test Day for Fedora Media Writer before Beta freeze" - it seems like that's definitely in progress and i think we can check in more detail during 'test day status' later in the meeting 15:10:12 sound OK? 15:10:30 yes 15:10:45 absolutely :) 15:11:34 ok :) 15:11:52 #info "sumantro and arvind (new RH interns) to set up Test Day for Fedora Media Writer before Beta freeze" - it seems like that's definitely in progress and i think we can check in more detail during 'test day status' later in the meeting, https://fedorahosted.org/fedora-qa/ticket/483 15:12:31 #info "adamw and pschindl to oversee the Test Day and make sure planning is well underway by next week (2016-04-11), otherwise step in" - as planning is moving forward we don't need to take over, just help out 15:13:13 "adamw to get dgilmore to chime in on the ticket with his thoughts" 15:13:17 this was about https://fedorahosted.org/rel-eng/ticket/6383 15:13:37 looks like dennis didn't comment yet, so i'd best poke him again 15:13:59 #action adamw to get dennis' thoughts on https://fedorahosted.org/rel-eng/ticket/6383 15:14:29 * linuxmodder late 15:14:34 #info "adamw to get dgilmore to chime in on [rel-eng #6383] with his thoughts" - didn't manage to do that this week 15:14:37 hi! 15:14:54 anything else for previous week follow up before we move onto this week's topics? 15:15:36 alrighty then 15:15:37 #topic Fedora 24 status 15:16:00 i just left this on the agenda as quick check-in topic, I don't have anything really specific. 15:16:07 good news I guess is that nightlies are mostly working again! 15:16:27 *yay* 15:16:41 finally, openQA tests are passing (more or less) 15:17:04 yep 15:17:44 are the tests passing, or is openQA failing? 15:17:55 so we should be looking to Beta tests at this point; remember with the new full-fat nightly composes there won't be so many 'candidate' composes, we may only start doing them after freeze, so testing the 'nominated' nightlies is important 15:18:01 * danofsatx wouldn't put it past adamw to 'modify' the code in that way 15:18:04 * satellit no soas since n.9 15:18:12 danofsatx: :P it's all for real, we promise 15:18:33 satellit: the lives are still subject to that live compose bug 15:18:46 satellit: so we still have the same 'lottery' effect - each day, some will fail 15:18:55 peter modified the .ks yesterday... 15:19:29 satellit: it could be that SoaS had a specific bug too, i didn't look at its logs 15:20:04 garretraziel: FWIW the few failures for 20160410.n.0 looked like timeouts to me, odd that we're hitting those but we'll see about 0411. the only real bug is the one the HDD kickstart test hit, next anaconda build should fix that 15:20:33 that occasional ext4 corruption bug that causes the image to blow up at some poine in boot needs more detailed logs at the time it's being created, i can't tell what's happening from the available info 15:20:54 #info Fedora 24 is broadly in good testable state right now, so please everyone help work on the current nightly validation tests 15:21:03 cmurf: dgilmore and bcl are both trying to figure it out, i think, but it's a tricky one 15:21:09 it's suspiciously coinciding with e2fsck discard 15:21:13 adamw: if we don't have enough tests failing, it means that we don't have enough tests 15:21:18 note - i might get a new nightly nominated for testing soon because the one that got nominated has no lives 15:21:23 garretraziel: =) true indeed 15:21:36 i wanted to get freeipa testing going for beta, but i got distracted by some other shiny objects... 15:21:43 jskladan said he had nothing to do, right? ;) 15:22:29 or we can just wait until I finish ARM tests :-D 15:23:12 that too ;) 15:23:23 alrighty, anything else for f24 status? 15:23:56 rpm -ql pencil 15:24:04 whoops, wrong window 15:24:27 alrighty, moving along 15:24:38 #topic Test Day status 15:24:46 so first off, we have a Test Day tomorrow, i18n 15:25:19 #info i18n Test Day is tomorrow: https://fedoraproject.org/wiki/Test_Day:2016-04-12_I18N_Test_Day 15:26:00 it looks like everything's ready to go there; I'm gonna tweak the page a bit to make it clearer to go to the test day app to find the tests, but otherwise it looks like it's all ready 15:26:19 #action adamw to tweak i18n test day page a bit so it's clearer that tests are found on the app page 15:26:36 otherwise, if people could come along tomorrow and help test that'd be great :) 15:27:12 it might be nice if there was more information on how to test the f24 changes i guess, oh well 15:27:27 any other notes on that test day? 15:28:31 a more content to how to test would be really helpful 15:28:44 for the i18n test day :) 15:28:57 * danofsatx would gladly help if he could speak a language other than Appalachain Hillbilly 15:28:58 on the results page - http://testdays.fedorainfracloud.org/events/4 - all the column titles are links 15:29:05 most of them are links to test cases 15:29:25 i always think we should make that clearer somehow, but I don't know how...for now I guess I'll just write it in the page 15:29:48 the only exceptions are the 'glibc locale subpackaging' and 'dnf-langpacks weakdeps' ones, those just link to the feature pages 15:31:08 alrighty 15:31:09 adamw, What exceptions does glibc have? 15:31:28 codonell: i mean it's an 'exception' in that the link on http://testdays.fedorainfracloud.org/events/4 doesn't point to a test case like most of the others 15:31:37 codonell: not really your problem, don't worry :) 15:31:44 Ah :-) 15:31:50 * kparal chuckles 15:31:53 adamw, If you need help from the fedora glibc team, just ask :- 15:31:55 ) 15:32:12 thanks! 15:32:31 * codonell stalks all glibc keywords ;-) 15:32:31 alrighty, so the other test day coming up is the live media writer / live USB creator one we mentioned in follow-up 15:32:35 codonell: i noticed =) 15:33:12 #info Live Media Writer (new liveusb-creator) Test Day planning for 2016-04-19 is underway, see https://fedorahosted.org/fedora-qa/ticket/483 and https://fedoraproject.org/wiki/Test_Day:2016-04-19_Fedora_Media_Writer 15:33:45 sumantro__: so i'd say the next step is to gather up test cases and write any necessary new ones 15:34:12 we have one general test case for the tool, https://fedoraproject.org/wiki/QA:Testcase_USB_stick_Live_luc 15:34:27 sure adamw 15:35:08 that links to instructions that describe the older luc interface, so it might be an idea to write a new test case covering the most basic use of the tool, using that test case as a guide 15:35:12 test that it makes media that boots on BIOS, UEFI, (ARM?) 15:35:17 test that it wipes that media 15:35:38 multiplied by linux, windows, and maybe os x 15:35:44 one way i'd suggest arranging things is to have a few test cases that cover the most important features, and try to get those run on all platforms 15:35:53 * satellit only f23 .isos at the moment in f24 LUC 15:35:59 If you're going to do that, maybe use dumpet and make some tests to run automatically before the manual stuff to see if it at least /looks/ sane? 15:36:04 then have some 'extended' test cases and just try to run those on any platform available 15:36:33 thanks cmurf , i am going to add all those test cases 15:36:50 the generated media images, that is 15:36:59 adamw can you point me to the resource for mac , i didnt find one 15:37:16 sumantro__: I believe mbriza said Mac version will not be ready for F24 15:37:16 sumantro__: as in, the actual tool? 15:37:36 the FMW tool 15:37:43 Fedora Media Writer . 15:38:07 best ask mbriza 15:38:19 pjones: i think we're kinda gonna have two things to look at: whether the tool actually writes sane images, but also actual app issues like, does it actually run, UI bugs, are the defaults sane, does the 'get list of images to write' stuff work, etc 15:38:34 *nod*. 15:38:40 yeah, ask mbriza; if we don't have a mac version all we can do is test on fedora and windows 15:39:08 I'm just suggesting that it's a relatively easy thing to do to keep you from getting to the expensive stuff in the bad failure cases. 15:39:18 (where expensive == manual labor) 15:39:29 yes, i will. I will be finishing the test for windows 8 and 10 tonight and then add the test cases tomorrow . :) 15:40:09 pjones: yeah, good idea - i'm not sure if it's in sumantro's wheelhouse or not, but noted. i do want to cover all the app issues too, though, because one of the things we need to be able to do here is tell fesco whether this thing is remotely likely to be ready for f24 or not 15:40:24 I'm sure you do, yes :) 15:41:03 sumantro__: i'll send you a mail with some ideas on how to organize the tests page, but i'd say that's the basic aim for now :) 15:41:16 I guess the stuff pjones noted would be best incorporated into the app itself, so maybe mention it to mbriza? 15:41:23 IIUIC 15:41:28 thanks adam :) 15:41:30 #action sumantro to work on test cases for the media writer Test Day, adamw to guide and help out 15:42:04 kparal: well, maybe. it has more to do with the expected output, but if all the cases are reasonably similar or distinguishable within the tool, then yeah. 15:42:13 sumantro__: I think just a single test case for Windows is fine (either generic, or Win10) 15:42:28 kparal: I think he means he's testing the tool on both himself 15:42:36 ah, ok 15:43:03 yes i will be testing it on both . 15:43:23 then write a generic set of test case for windows 15:44:33 sumantro__: the way i like to do this kind of 'test everything in multiple cases' is to write generic test cases, then have the test results table handle the environments 15:44:54 so we'd have tests A, B and C and the test table would have the tests as columns and the environments as rows 15:45:25 yeah if there's something weirdly unique about a platform then that could be a note in the test case, but doesn't mean separate testcases need written 15:45:27 sumantro__: so like http://fpaste.org/354115/89522146/ 15:45:45 sumantro__: we can set it up like that in the test days app, i'll explain in more detail in a mail :) 15:46:02 got it adamw! 15:46:28 alrighty, i think we know where we're going with that then 15:46:30 moving along to... 15:46:33 #topic Open Floor 15:46:40 so i'm gonna steal the mic here briefly 15:47:17 i have a couple of pet projects to mention...since qa-devel meetings happen way too early and this week's got cancelled anyhow... 15:47:31 * nirik has something to mention at some point if there's time after adamw. 15:48:10 12 minutes, speed it up, adamw :) 15:48:17 as i mentioned on the list I wrote a thing for finding nightly images, which you can find at https://www.happyassassin.net/nightlies.html . when you just want to find the latest Workstation live or whatever for Rawhide or Branched, you can use that. i'm hoping we can make it something 'official' and move it to fp.o, i'll talk to folks about that. code is at https://pagure.io/fedora_nightlies , feel free to send issues and prs and stuff 15:48:33 #info nightly finder app at https://www.happyassassin.net/nightlies.html and https://pagure.io/fedora_nightlies , yay 15:48:53 greatly appreciated, btw. we were missing something like this 15:49:11 dammit jim! 15:49:12 i also made a thing which runs anaconda's kickstart-tests - https://github.com/rhinstaller/kickstart-tests - in openQA. now i'm kinda looking for feedback on whether to move forward with that or not 15:49:36 that thread is on anaconda-devel-list: https://www.redhat.com/archives/anaconda-devel-list/2016-April/msg00001.html 15:49:43 if folks could chime in with their thoughts i'd really appreciate it 15:49:43 adamw: if you wanted to create projects/mirror repos in Phab for it, just let me know 15:49:52 I think that it's great idea 15:49:58 kparal: i think i'm kinda ok with it being in pagure for now 15:50:10 with ks-tests 15:50:14 garretraziel: there's more detail to the question in the mail =) not just 'should we do it?' but 'for whom?' and 'where?' and 'how exactly?' 15:50:31 kparal: since it's not really a "QA" project, i'm hoping to offload it onto the infra or apps folks at some point 15:50:40 adamw: sneaky! 15:50:53 #info thoughts on openQA running of kickstart-tests welcome at https://www.redhat.com/archives/anaconda-devel-list/2016-April/msg00001.html 15:51:03 alright, go nirik =) 15:51:26 * kparal read the anaconda thread, but does not have enough insight to guess what the best approach is 15:52:17 kparal: well, it would be useful to have a taskotron expert just look at the kickstart-tests and think about how well they'd fit into taskotron 15:52:44 so, there's going to be a storage upgrade (likely/possibly next tuesday, the 19th). For that they need to throttle our io for a few days before/after... so we are possibly going to lock down rsync access to dl.fp.o to tier1 mirrors to help that out. How much is that likely to impact all you folks? 15:52:49 kparal: i know openQA pretty well so it was natural for me to look at running them there, but i don't know taskotron as well at all, so it's hard for me to answer how much more sense it makes to do them there and how much work it'd be; maybe together we can get the answer =) 15:53:42 nirik: um. fedfind finds images for non-Pungi 4 composes by rsync scraping, so it would be affected 15:53:50 adamw: yeah, that I can do. will try 15:54:28 nirik: i can't quite think off the top of my head if that breaks anything important...i'll have to think on it 15:54:30 ok. The alternative would be that it would be slow for everyone and tier1 mirrors couldn't get synced. ;( So, not sure there's a great answer... 15:54:46 nirik: possible workaround is to flip it over to using a different server for the duration, but then i'd have to get all relevant systems updated. 15:54:52 I just wanted to bring it up in case anyone saw any problems we could mitigate. 15:54:58 * nirik just likes typing mitigate. 15:55:02 nirik: brno office syncs fedora-alt candidate composes from dl.fp.o. will it continue to work? 15:55:03 nirik: how long is this gonna take>? 15:55:21 adamw: download-ib01 should hopefully be synced. 15:55:30 kparal: yeah, I think globalsync would keep working... 15:55:43 nirik: ok, great 15:56:01 they were talking about a few days before... so monday? sunday? upgrade on tuesday, possibly re-enable things tuesday night? 15:56:06 if it's going to be for a few hours or even up to a couple of days i think we can just wait it out, longer than that i might try to figure out if anything important will stop working and deal with it somehow 15:56:26 a couple of days is not a problem 15:56:32 the fact that mirror scraping isn't used for pungi 4 composes means *most* stuff should be fine, i'm just trying to figure out if there's any little cases i'm forgetting 15:56:41 also, we have come up with a much better mirroring setup we hope to have in place... might look at what it would take for fedfind to use that setup as soon as we have it in place. 15:56:51 (also i sync my personal mirror from dl.fp.o but i can switch that over easily enough =)) 15:57:07 nirik: yeah, please do send me the details. any time stuff moves around, fedfind needs to know =) 15:57:10 basically: make a fullfilelist.something that has file 15:57:28 so, you can download and look at what changed in that file instead of doing recursive rsync 15:57:43 and then download just those files that changed since the last time you synced 15:57:46 nirik: will you have some wrapper over rsync to automate that? 15:58:03 kparal: That's what we were going to write once the file list exists. 15:58:07 yeah, hopefully a simple shell script 15:58:23 nirik: https://www.happyassassin.net/cgit/fedfind/tree/fedfind/release.py#n1186 is what fedfind uses rsync for. 15:58:25 sounds awesome and I considered that many times when fighting with globalsync 15:58:27 We also need some way to communicate that something should be hardlinked. 15:58:37 can't we use the repodata files? 15:58:42 however I'm afraid globalsync will not be able to use that system 15:58:52 danofsatx: For a mirror? I doubt it. 15:59:00 Not all of the content there is in any kind of repository. 15:59:10 kparal: I fear the same. ;( 15:59:20 oh, ok...I need to read closer. sorry. 15:59:27 * danofsatx goes back to his corner. 15:59:30 kparal: But if even some of the mirrors stop doing full syncs then it helps everyone. 15:59:41 * kparal agrees 15:59:49 Over eight million files for a full copy now. 16:00:03 ok, we're about at time 16:00:08 anything else? 16:00:25 we are doing about 2,000IOPS read, 2,000IOPS write and 25,000 for metadata. ;( 16:00:26 Oops, I thought I was in a completely different channel. 16:01:09 let's say blocker meeting will start in 5 mins in #fedora-blocker-review 16:01:13 so folks can grab a drink or whatever 16:01:51 tibbs|w: you are in a twisty maze of irc channels. 16:02:03 #info rsync access to dl.fedoraproject.org may be blocked for a day or two around 2016-04-19, please discuss any issues with nirik 16:02:14 thanks for coming everyone! 16:02:17 #endmeeting