15:00:04 #startmeeting Fedora QA Meeting 15:00:04 Meeting started Mon Sep 20 15:00:04 2010 UTC. The chair is jlaska. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:04 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:09 #meetingname fedora-qa 15:00:09 The meeting name has been set to 'fedora-qa' 15:00:21 #topic Gathering carbon units 15:00:55 alright ... show of limbs, who do we have lurking? 15:01:39 * Viking-Ice ... 15:01:46 Hi Viking-Ice 15:01:50 * hircus ! 15:02:06 * kparal here 15:02:19 welcome hircus + kparal 15:02:38 * Oxf13 15:02:51 * wwoods 15:03:35 Oxf13: welcome from CEST 15:04:11 wwoods arrived in a shoe? 15:04:42 do we have adamw lurking in the shadows? 15:04:51 if not, we'll save the F-14-Beta update until the end 15:05:03 * dgilmore is around 15:05:12 dgilmore: hi 15:05:29 okay, let's get started ... 15:05:31 morning people :-) 15:05:45 saccia: howdy 15:05:51 #topic Previous Meeting follow-up 15:06:09 This will be quick ... since I see no action items from last weeks meetbot minutes 15:06:28 shout if I missed something, otherwise I'll move on 15:07:06 alright ... skipping F-14-Beta status for now ... we'll come back to that when adamw has arrived 15:07:13 #topic Fedora 14 Test Days 15:07:50 The installer keymap/lang event last week seems to have been quite successful in terms of bugs 15:08:13 looks like good participation on the wiki, that's always positive 15:08:25 #info installer keymap/lang test day recap - http://lists.fedoraproject.org/pipermail/test-announce/2010-September/000133.html 15:08:43 in case you aren't subscribed to test@ or test-announce@ see the link above for the recap from Igor 15:09:03 next up ... Virtualization this Thursday 15:09:22 #info Sep 23 - Virtualization Test Day - https://fedoraproject.org/wiki/Test_Day:2010-09-23_Virtualization 15:09:36 jforbes: nice, the wiki is coming along 15:10:09 jforbes: anything we can do to help keep this event focused? 15:10:40 jlaska: nope, getting the last of it up today, and going to send out an email to fedora-vert 15:10:45 err virt even :) 15:11:21 heh ... okay, adamw or I (anyone else encouraged) can hit up Fedora planet with a post about the event as well 15:11:53 Yeah, I will do that too 15:11:56 jforbes: were there specific features you wanted covered for this event, or is this more of a general virt testing theme? 15:12:12 more general, but spice will be covered 15:12:13 * jlaska always worried that we try to fit too much into a single virt test day 15:12:42 hmm, wonder if it would make sense to theme it around spice? 15:12:55 some of the packages listed in "Update your machine" are not the latest as of today. Perhaps the page can be updated? esp. for the kernel version 15:12:58 Not sure how much of an audience spice will have 15:13:27 hircus: The page will be updated on wednesday, as there will be updates between now and then 15:13:28 jforbes: is that hardware specific, or just new enough that we wouldn't have a lot of participation 15:14:27 jlaska: both limited audience due to desktop nature, and lack of libvirt/virt-manager integration will make it less user/tester friendly 15:14:39 jforbes: aah, I see (re: the integration portion) 15:14:59 But we also have people who will come to test that who would not otherwise participate, it does have a lot of interest 15:15:31 nice! 15:16:00 jforbes: well, shout if you need help/guidance on the wiki setup (shouting in #fedora-qa helps) 15:16:14 jlaska: will do, thanks 15:16:25 following virtualization, we have the always well attended ... 15:16:39 #info Sept 28-30 - Graphics test week 15:16:56 Adam has wiki pages in place already, so it looks like things are in good shape 15:17:08 I'm sure we'll know more later in the week 15:17:22 that's all I had on test days 15:17:44 next up ... autoqa 15:17:57 #topic AutoQA Package Update Acceptance 15:18:18 wwoods: do you want to kick things off? 15:18:37 * kparal noticed a flood of new patches lately 15:19:03 kparal: you can blame me for some of those 15:19:42 how much can happen in just a few days 15:20:01 until wwoods wake up, I'll just note that I presented AutoQA in FUDCon this weekend 15:20:08 you can see the slides here: https://fedoraproject.org/wiki/QA:Presentations 15:20:27 unfortunately there are no recordings of the talk 15:20:28 well, I pushed out autoqa-0.4.1-1, and then tried to fix several quick bugs I found so you had a mostly functional autoqa for your talk (autoqa-0.4.2-1) 15:20:45 * wwoods appears 15:20:54 kparal: nice blog posts on the subject as well 15:21:02 it was a nice talk 15:21:20 there was some genuine interest in the audience, and I apologize for hijacking the talk by answering some of the questions :) 15:21:35 Oxf13: thank you for that 15:22:00 * kparal leaves mic to wwoods 15:22:31 so, work on depcheck continues, albeit slowly 15:22:34 #info kparal gave AutoQA presentation at FUDCon Zurich - https://fedoraproject.org/wiki/QA:Presentations 15:23:19 more interesting, we merged the 'multitests' branch into master, so now we should have proper support for multihook tests and things 15:23:46 but that's all kparal and jskladan's work, so hooray for them 15:23:55 which makes me wonder whether the test result mails should prefix the subject with the hookname? But that's a detail for autoqa-devel@ 15:24:08 jlaska: is that stuff in 0.4.2 or are we still shooting for an 0.4.5 (or similar) release for that? 15:24:28 jlaska: good point - and yes, let's discuss that on the list 15:24:30 wwoods: bodhi should start utilizing the new pending tags today... rolling the code out as we speak. 15:24:40 wwoods: 0.4.2 is the current release ... mainly due to my failure to read the list before tagging 0.4.1 15:25:03 lmacken: excellent! jskladan gave me some code demonstrating how we can use mash with or without a tag 15:25:23 so I'll be writing some unittests to prove the mashing / multilib stuff works as expected in depcheck 15:25:28 wwoods: oh, interesting. 15:25:28 and then soon we'll actually wire it up 15:25:38 sounds good 15:26:07 #info lmacken noted that bodhi should start utilizing new 'pending' tags today 15:28:16 wwoods: anything other updates? 15:28:24 oh - I updated the Updates Lessons page 15:28:30 with information about the nss-softokn problem 15:28:43 https://fedoraproject.org/wiki/Updates_Lessons 15:28:46 wwoods: oh right, thank you ... that one was a bit complex 15:29:15 wwoods: what's the next milestone for the depcheck test? 15:29:28 I saw a ticket update from jskladan last week that he got something working with mash 15:30:37 kparal: anything we need to get on the radar based on the feedback from your talk? 15:31:02 jlaska: right - next milestones are getting depcheck using mash - to make sure its input matches what bodhi would push 15:31:30 and making depcheck do an RPM test transaction - like yum does - to make sure its test matches what yum would do on people's systems 15:31:41 this is for capturing file conflicts? 15:32:01 jlaska: I received comments mainly on whitelisting rpmlint output and using beaker. the first thing is being solved on our ML, but I had still no time to read through it. the second thing can be discussed later, I think it's still not fully opensource yet 15:32:05 jlaska: one suggestion I recall from the talk is having Koji integration for AutoQA 15:32:42 hircus: "integration"? 15:33:07 jlaska: yes 15:33:11 I think hircus means what I had in my slides. showing autoqa results right next to the newly built packages in Koji page 15:33:25 ah, integration in the web frontend. yes, that would be cool 15:33:38 hircus: yes, we have it in our plans :) 15:33:55 which we are hand-waving and referring to as "resultsdb" for now 15:34:02 another aspect that some people were asking is to do the local builds inside mock / koji 15:34:05 resultsdb - the magic test results web service 15:34:27 ah, yes. I wasn't able to answer those mock questions 15:34:36 that's... not really a QA problem 15:34:54 #info wwoods and jskladan continue to work on depcheck and integration with mash 15:35:19 hircus: yeah, what wwoods said. Was this specific to something about QA, or just general mock/koji? 15:36:14 alright ... anything else on the autoqa front? 15:36:19 if not ... we'll dive into F-14-Beta 15:36:23 (head first) 15:36:38 it's QA related, in a way -- I could imagine autoqa tests would be more reproducible if you can control the local environment 15:37:39 hircus: autoqa tests are designed to be easy to run and not require a lot of setup/prep 15:37:55 fair enough 15:38:10 if there are some specifics that are painful to work with ... do speak up though 15:38:25 but for now, let's take that to autoqa-devel@ for further discussion 15:38:39 #topic F-14 Beta Testing 15:38:58 adamw isn't around, so I'll go off of current status based on F14Beta and several email threads 15:39:24 RC#2 contains the following open F14Beta blocker bugs .. 15:39:27 635396 - ConsoleKit session not properly opened (xfce) 15:39:27 635395 - ConsoleKit session not properly opened (lxde) 15:39:27 635332 - FC14 Beta RC2 i386 netinstall fails to boot 15:39:28 635330 - Boot menu is missing in F14 Beta RC2 install discs 15:39:50 the last is clearly a blocker, given the recently added criteria around artwork 15:40:14 I'm still trying to gather more information around the RC2 netinstall failure, I've not had any problems with netinstalls so far 15:40:28 I'm really interested in a retrospective as to why 396 and 395 were found so late 15:40:40 And the first 2 issues (ConsoleKit) are fairly major for both spins 15:40:41 jlaska: I experience this with more than netinstall.... I've updated the bug with what happened 15:40:48 Oxf13: likewise, /me adds 15:41:01 the "Where to install bootloader" screen doesn't appear even though I have two drives 15:41:09 * nirik is having trouble duplicating 635396 15:41:44 I also might consider adding https://bugzilla.redhat.com/show_bug.cgi?id=622694 as a beta blocker too if it seems fair 15:42:12 saccia: do you get the screen where you get to pick the 2 drives 15:42:17 for install/boot devices? 15:42:27 and it has the radio button to choose which drive to install the bootloader 15:42:36 jlaska: no... I took "screen shots" that show it appears in alpha but not in RC1/RC2 15:43:35 saccia: I believe that screen depends on your drive and partition selections, but let's continue to diagnose that after the meeting in #fedora-qa 15:44:03 with regards to the mouse click, I'm still not able to reproduce 15:44:12 but it seems that two of you are seeing this problem on test@ 15:44:39 jlaska: do we have all the info on the setups? video cards, type of mouse and how its connected? 15:45:08 dgilmore: not in the bug mentioned above 15:45:31 jlaska: seems like we need to know more about the hardware people are having issues on 15:45:32 the bug mentions another issue, 632431, which I'd consider a feature request 15:45:39 dgilmore: of course, agreed 15:46:02 so, summary ... F14Beta is accurate for what we know right now 15:46:27 and there are 2 issues under discussion (those noted by saccia) to determine their impact to users 15:46:27 dgilmore: if you can add something in the bug detailing how to get whatever logs/info you need I'll be happy to get that info 15:46:31 jlaska: at the leastw e are going to need anaconda and likely a ConsoleKit update 15:46:36 dgilmore: yeah 15:46:44 on top of this ... we need to complete the planned testing 15:46:54 #link http://fedoraproject.org/wiki/Test_Results:Current_Installation_Test 15:47:01 #link http://fedoraproject.org/wiki/Test_Results:Current_Desktop_Test 15:47:36 with the late start and respins ... we're obviously short on test time 15:48:04 jlaska: go/nogo is wednesday? 15:48:04 I think sticking with the phrase 'at risk' also accurately describes QA testing for F-14-Beta 15:48:21 dgilmore: you got it ... Wed 2010-09-22 15:48:37 #info 2010-09-22 - Fedora 14 Beta Go/No-Go Meeting (17:00 EST) 15:48:39 jlaska: ok so we really need to get rc3 done today 15:48:41 time == crunch 15:49:11 dgilmore: yeah, by end of day today at least 15:49:16 jlaska: do you feell that testing rc2 livecds for desktop is ok? 15:49:31 dgilmore: sorry, can you rephrase? 15:50:04 jlaska: should i try do some testing with the livecds today while waiting for fixes? or best to wait for rc3 15:50:20 dgilmore: oh I see ... I'd go ahead and test anyway 15:50:30 I'm doing my best to fill out results in the install matrix using RC2 15:51:04 any other questions/concerns around F-14-Beta? 15:51:05 some tests can be marked as fail just because of the missing boot menu 15:51:14 memtest for example 15:51:23 robatino: good point, I haven't marked those as fail yet 15:51:40 I wonder if we need a new indicator 15:51:40 * jlaska trying to knock out the hardware, partitioning, updates and traceback cases 15:51:46 blocked vs pass/fail 15:52:40 Oxf13: meh ... no real opinion there ... I'm happy to update https://fedoraproject.org/wiki/Template:Result if there is overwhelming desire for blocked 15:53:11 not sure if it's overwhelming, just wondering if a blocked test means anything different to us than a failed test 15:53:36 right now, I haven't found a big need to distinguish between the two 15:53:42 some sort of relational test matrix would help 15:53:51 you're right, it certainly has different meaning 15:54:10 The has helped for re-using existing bugs/failures for test results 15:54:29 so at least we have a mostly unique list of bugs in the References section 15:54:58 alright ... any other F-14-Beta thoughts? 15:55:00 k 15:55:22 if not ... 15:55:29 #topic Open Discussion - 15:55:48 thoughts/comments/concerns/hicu 15:56:40 if not ... let's close this out in 2 minutes and get back to F-14-Beta 15:56:42 haikus are easy; But sometimes they don't make sense; Refrigerator 15:56:55 bz#627789 15:57:12 Oxf13: thank you for proper spelling! :) 15:57:22 Southern_Gentlem: care to join #fedora-devel and talk about your respin issue? 15:57:30 It didn't make it into f14-beta-branch, and I'm hitting it when I use livecd tools to make a usb stick with the DVD iso. 15:58:10 I probably shouldn't have closed it as fixed in rawhide. But I also don't see a large number of other reports on it. 15:58:23 bcl: only 1 dup so far 15:58:38 " 15:58:38 "Problem was a race between unmounting the iso and deactivating the loopback. 15:58:41 mount can now handle mounting iso's with loop without needing losetup." 15:59:08 bcl: is this a fix we need to consider for F14Beta? 15:59:41 I think so, but that's because I keep hitting it with my usage :) 15:59:48 haiku Fedora 14 / Is still very green and new / More tests must we do 15:59:50 #info 627789 - Error setting up repository - 16, Device busy 16:00:15 saccia: kudos :) 16:00:41 bcl: I'll ping you in #anaconda afterwards to understand more about this 16:00:46 ok 16:01:03 alright gang ... thanks for your time. I'm calling end of meeting 16:01:13 happy testing :) 16:01:18 thanks jlaska! 16:01:18 #endmeeting