16:01:32 #startmeeting f20alpha-blocker-review-3 16:01:32 Meeting started Wed Sep 4 16:01:32 2013 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:32 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:01:32 #meetingname f20alpha-blocker-review-3 16:01:32 The meeting name has been set to 'f20alpha-blocker-review-3' 16:01:32 #topic Roll Call 16:01:57 * jreznik is here 16:02:00 * satellit_e listening 16:02:11 * pwhalen is here 16:03:54 * pschindl is here 16:03:56 * tflink waits a few more minutes to see if we get a couple more people 16:04:13 any volunteers for secretary duty? 16:07:03 * tflink takes that as a no 16:07:31 well, lets get this party started. I guess I'll do secretarialization post-meeting 16:08:21 * kparal arrives late 16:08:32 #topic Introduction 16:08:36 #chair kparal 16:08:36 Current chairs: kparal tflink 16:08:44 Why are we here? 16:08:44 #info Our purpose in this meeting is to review proposed blocker and nice-to-have bugs and decide whether to accept them, and to monitor the progress of fixing existing accepted blocker and freeze exception bugs. 16:08:51 #info We'll be following the process outlined at: 16:08:51 #link https://fedoraproject.org/wiki/QA:SOP_Blocker_Bug_Meeting 16:08:58 #info The bugs up for review today are available at: 16:08:59 #link http://qa.fedoraproject.org/blockerbugs/current 16:09:05 #info Up for review today, we have: 16:09:12 #info 8 Proposed Blockers 16:09:12 #info 5 Accepted Blockers 16:09:12 #info 0 Proposed Freeze Exceptions 16:09:12 #info 3 Accepted Freeze Exceptions 16:09:29 if there are no objections, we'll start with the proposed blockers 16:10:10 #topic (1000703) TypeError: not enough arguments for format string 16:10:13 #link https://bugzilla.redhat.com/show_bug.cgi?id=1000703 16:10:15 #info Proposed Blocker, anaconda, MODIFIED 16:11:07 am I the secretary today, or was there another volunteer? 16:11:52 kparal: no other volunteer, if you could take care of it, I'd appreciate that 16:12:05 so, this sounds like "custom partitioning screen is broken" to me 16:12:19 tflink: ok 16:13:02 this isn't really an alpha blocker 16:13:09 but it would be nice if it didn't crash right away 16:13:44 hmm weren't all installer crashes blockers? 16:13:55 Viking-Ice: no 16:14:19 ( as in visible/hard crashes ) 16:14:44 not at alpha, I don't think 16:14:46 * kparal failed to find appropriate criterion, Alpha speaks about guided part only 16:15:10 -1 blocker +1 fe 16:15:11 -1/+1 16:15:23 that sounds more like final 16:15:27 but -1/+1 16:15:48 -1/+1 as looking on criteria 16:16:25 proposed #agreed 1000703 - RejectedBlocker AcceptedFreezeException - This doesn't violate any F20 alpha release critera and thus does not qualify as a release blocking bug. However, a tested fix would be considered past freeze. 16:16:51 ack 16:17:00 btw, custom part is Beta? 16:17:02 ack 16:17:28 kparal: right, beta http://fedoraproject.org/wiki/Fedora_20_Beta_Release_Criteria#Custom_partitioning 16:17:31 ack 16:17:59 do we want to take it as a beta blocker, then? 16:18:09 I'll just re-propose 16:18:16 it's already modified anyway 16:18:16 k, thanks 16:18:22 #agreed 1000703 - RejectedBlocker AcceptedFreezeException - This doesn't violate any F20 alpha release critera and thus does not qualify as a release blocking bug. However, a tested fix would be considered past freeze. 16:18:34 #topic (1002811) Anaconda is unable to handle dependcy problems without crashing 16:18:37 #link https://bugzilla.redhat.com/show_bug.cgi?id=1002811 16:18:40 #info Proposed Blocker, anaconda, NEW 16:19:20 -1/-1 this is a "feature" request 16:21:12 other thoughts? 16:21:31 we still need info on which package manager the installer is using dnf/yum/what ever the desktop team is working on and what gnome 3 upstream is also working on 16:21:33 so the bug report is about the firefox bug? 16:21:42 well, I expect for "final" release we would have all dependencies issues fixed, right 16:21:59 dependency should be fixed at alpha 16:22:14 no, that's just what steve was hitting when attempting to test 16:22:16 * spstarr_work is here 16:22:19 just... working :) 16:22:28 actually if we can test this in "testing transaction" phase, I don't believe the installer should fail after adjusting the partitions 16:22:31 this is a request to get anaconda to ignore dep problems w/o stopping install 16:22:50 there's stress on _after_ 16:22:51 so it's not blocking alpha if it crashes as there should not be any issue at all - so really, it's feature request 16:23:09 hmm, after... 16:23:22 jreznik: I think the firefox deps are not broken, it's just not installable 16:23:28 due to script failures 16:23:34 yeah, the pkg install just fails 16:23:44 so I don't know whether we cover that by other Alpha criteria 16:23:54 but maybe that's just beauracracy 16:23:59 this isn't the firefox issue 16:24:04 * kparal never knows how to spell that word 16:24:10 * tflink neither 16:24:14 this is dep problem right 16:24:34 this is about anaconda not continuing the install if it hits transaction problems during install 16:24:53 "There must be no errors in any package on the release-blocking images which cause the package to fail to install. " 16:24:57 yeah well that's feature as you mentioned -1/-1 16:25:00 actually this cover it well 16:25:07 not really 16:25:22 this isn't about an error in any package, it's about how anaconda handles the error 16:25:53 the mention of 1003691 is an anecdote explaining why farther testing couldn't be done 16:26:08 so it's possible to continue with the transaction, and anaconda chooses to exit? 16:26:26 not as sure about that 16:26:38 I dont think that's possible 16:27:05 as in to "ignore" problem package and "proceed" with install 16:27:26 even if it is possible, is it really desirable? 16:27:41 yup 16:27:43 ok. I think the bug is missing a lot of information. it says "anaconda is crashing" but doesn't provide any log files. we don't know if it just prints out an error before install, or during install, or anything. also, we cover a lot of package problems with other criteria 16:28:07 well we have 2x -1/-1 16:28:09 so, let's say "add more info or rejected" 16:28:10 already 16:28:17 it's a feature request to change anaconda's behavior 16:28:37 * tflink only saw 1 vote 16:28:43 oh, i see the second 16:29:44 proposed #agreed 1002811 - RejectedBlocker RejectedFreezeException - This doesn't violate any F20 alpha release criterion and is a request for anaconda behavior change and thus doesn't qualify as a release blocking bug for F20 alpha. 16:29:48 I'm -1 on the basis that we can cover the breakage by other criteria. I don't think that anaconda should crash during the installation because of broken package, if it can test it in advance. but that wouldn't be Alpha blocker, probably 16:29:51 ack/nak/patch? 16:30:21 ack 16:30:29 ack 16:30:44 I'll add a comment 16:30:45 ack 16:30:46 #agreed 1002811 - RejectedBlocker RejectedFreezeException - This doesn't violate any F20 alpha release criterion and is a request for anaconda behavior change and thus doesn't qualify as a release blocking bug for F20 alpha. 16:31:10 kparal: yeah, the actual breakage (where it exists) would be covered by other criteria 16:31:28 not sure we want to take something like this as a blocker w/o talking w/ anaconda devs first 16:31:39 #topic (1003229) TC2 Net installer fails with Error 256 16:31:40 #link https://bugzilla.redhat.com/show_bug.cgi?id=1003229 16:31:40 #info Proposed Blocker, anaconda, NEW 16:32:04 * tflink couldn't reproduce the issue, is pretty sure it's a mirroring problem 16:32:05 -1/-1 16:33:10 hmm, I wonder whether it can be connected to 1003637 16:33:11 Anaconda should better handle that but yeah -1/-1 16:33:31 * jreznik is back 16:33:56 -1 because there are no logs 16:34:10 it's strange feeling to see all bugs starting with 1 16:34:32 the anaconda team should have triage this and put it into needinfo status 16:34:35 kparal: it's not an anaconda issue 16:35:01 no mirrors to try... 16:35:01 I'm 90% sure this is a mirror problem - I've hit it more than once and found a broken mirror to be the cause 16:35:02 tflink: no, but I suspect some yum changes 16:35:44 yeah, dgilmore has been having all kinds of fun with yum and f20 composes 16:35:54 I see -2 16:36:20 * jreznik is -1 but would be nice to try to get opinion from yum guys 16:36:25 do we have any info if the anaconda team is planing on changing the package management ( yum vs dnf ) they use ? 16:36:45 proposed #agreed 1003229 - RejectedBlocker - This appears to be a mirror issue and does not violate any F20 alpha release criteria, thus rejected as a release blocking bug for f20 alpha. 16:36:55 this _isn't_ a yum/dnf bug 16:37:03 ack 16:37:19 anaconda _is_ still using yum, doesn't it? 16:37:23 the error message is a bit weird, sure but why should we expect yum to deal with a busted mirror 16:37:30 ack 16:37:49 kparal, I would think they would not switch unless we make an "official" switch 16:38:16 moar ack/nak/patch? 16:38:32 ack 16:38:34 #agreed 1003229 - RejectedBlocker - This appears to be a mirror issue and does not violate any F20 alpha release criteria, thus rejected as a release blocking bug for f20 alpha. 16:38:42 #topic (994180) Boot-time LUKS passphrase input *always* defaults to en-us 16:38:45 #link https://bugzilla.redhat.com/show_bug.cgi?id=994180 16:38:48 #info Proposed Blocker, anaconda, NEW 16:38:51 * Martix is here 16:38:53 anaconda has the start of dnf support 16:39:15 in other words broken implementation 16:39:16 dgilmore: but not used by default I suppose 16:39:48 we need to get this cleared up from the anaconda team 16:39:55 jreznik: not yet 16:40:00 seriously they will no be shipping both methods 16:40:25 that would be plain stupid 16:40:39 Martix: welcome to the party 16:40:57 so, lbrabec tested this one thoroughly 16:41:08 and I'm +1 blocker 16:41:22 it sounds like the source of the issue has been found but there is a bit of disagreement on how to fix it 16:41:25 it's very simply. if you encrypt your disk with non-us letters, you have no way of decrypting it 16:41:37 but live only, no? 16:41:38 +1 16:41:43 or am I reading too fast 16:41:45 kparal: I can confirm this bug 16:41:47 if you use only ascii characters, you can work around by typing as with us keyboard 16:42:44 tflink: lbrabec tested with Live only, I guess. I'm not sure if it affects also DVD, but I expect so 16:43:04 anaconda simply doesn't regenerate initramfs after providing values to /etc/vconsole and others 16:43:23 so it should not be affected by Live/nonLive environment 16:44:07 anaconda does not due that ( which causes many bugs ) 16:44:11 it should 16:44:14 Vratislav Podzimek doesn't like generating initramfs for the second time, but that's a technical argument between him and Harald 16:44:26 not really 16:44:32 it should period 16:44:47 for the purposes of this discussion that's not important 16:44:49 Vratislav just has to suck it up and accept that fact 16:44:53 and it ended up with hostonly bashing... 16:45:23 +1 blocker, go on 16:45:33 the impact is that all people encrypting disk with non-ascii chars don't have any easy workaround 16:46:06 which violates the criterion sufficiently, I believe 16:46:16 proposed #agreed 994180 - AcceptedBlocker - Violates the following F20 alpha release criterion for systems with a non-us keyboard and encrypted disks: "A working mechanism to create a user account must be clearly presented during installation and/or first boot of the installed system." 16:46:22 ack 16:46:30 patch 16:46:37 "In all of the above cases, if any system partitions were encrypted as part of the installation, the boot process must prompt for the passphrase(s) and correctly unlock the partition(s) when provided with the correct passphrase(s)." 16:46:44 that's better criterion, no? 16:46:48 ah yeas 16:46:51 it is 16:46:57 no, it's not a criterion 16:47:10 it's a qualifier to the criterion, in my mind 16:47:31 alright, I'll just append it :) 16:47:33 ack 16:47:53 yeah, I don't know why I spend time formatting these proposals, they get changed when in bugzilla anyways 16:48:49 ack 16:48:50 the criterion applies is a bit weird 16:48:52 moar ack/nak/patch? or is everyone waiting for a change in proposal? 16:48:53 *applied 16:48:53 ack 16:49:00 kparal: applies? 16:49:14 you know, firstboot tool != encrypted disks 16:49:41 but that's not really important 16:49:47 firstboot tool is used as an indication of a successful install 16:50:00 ok ok 16:50:04 you can't get to that milestone of initial boot if you can't unlock the disk 16:50:08 potato potato 16:50:09 let's accept :) 16:50:20 #agreed 994180 - AcceptedBlocker - Violates the following F20 alpha release criterion for systems with a non-us keyboard and encrypted disks: "A working mechanism to create a user account must be clearly presented during installation and/or first boot of the installed system." 16:50:34 #topic (1004323) Apper: "You have failed to provide correct authentication." while checking/installing updates 16:50:38 #link https://bugzilla.redhat.com/show_bug.cgi?id=1004323 16:50:39 let's move on decided blocker is still an blocker regardless how it's phrased 16:50:40 #info Proposed Blocker, apper, NEW 16:51:07 +1 16:51:18 is apper our default graphical package manager 16:51:20 ;) 16:51:24 for kde, yes 16:51:33 kde aint default ;) 16:51:36 but yeah +1 16:51:44 but is a release blocking desktop 16:52:30 is this gpg keys problem? 16:52:50 proposed #agreed 1004323 - AcceptedBlocker - Violates the following F20 alpha release criterion for KDE: "The installed system must be able to download and install updates with yum and with the default graphical package manager in all release-blocking desktops." 16:52:56 ack 16:52:56 doesn't sound like it to me 16:53:01 ack 16:53:29 ack 16:53:38 ack 16:53:40 #agreed 1004323 - AcceptedBlocker - Violates the following F20 alpha release criterion for KDE: "The installed system must be able to download and install updates with yum and with the default graphical package manager in all release-blocking desktops." 16:53:45 #topic (1003691) firefox pretrans scriptlet issues 16:53:46 #link https://bugzilla.redhat.com/show_bug.cgi?id=1003691 16:53:46 #info Proposed Blocker, firefox, ON_QA 16:53:51 +1 16:54:04 There must be no errors in any package on the release-blocking images which cause the package to fail to install. 16:54:47 proposed #agreed 1003691 - AcceptedBlocker - Violates the following F20 alpha release criterion for firefox, which is part of the default gnome install: "There must be no errors in any package on the release-blocking images which cause the package to fail to install." 16:55:22 +1 16:55:29 ack 16:55:34 kinda autoblocker 16:55:35 -ack 16:55:37 ack 16:55:38 mean ack 16:55:48 #agreed 1003691 - AcceptedBlocker - Violates the following F20 alpha release criterion for firefox, which is part of the default gnome install: "There must be no errors in any package on the release-blocking images which cause the package to fail to install." 16:55:50 * jreznik is moving to another meeting, will be back soon 16:56:04 #topic (1002737) initial-setup-graphical.service does not run - TypeError: Argument 1 does not allow None as a value 16:56:07 #link https://bugzilla.redhat.com/show_bug.cgi?id=1002737 16:56:09 #info Proposed Blocker, initial-setup, NEW 16:56:57 +1 16:56:59 * tflink glares at pwhalen for not including a violated criterion even though this one is rather obvious 16:57:01 this one prevents initial-setup from running on arm images with desktop 16:57:24 tflink, sorry, will do going forward 16:57:26 "A system installed with a release-blocking desktop must boot to a log in screen where it is possible to log in to a working desktop using a user account created during installation or a 'first boot' utility."? 16:57:37 pwhalen: does minimal image include initial-setup as well? 16:57:49 * kparal just curious 16:57:50 it includes a text version, which runs.. 16:57:54 ok 16:58:00 +1 16:58:05 +1 16:58:06 there is initial-setup-{graphical,text} 16:58:10 +1 16:58:36 proposed #agreed 1002737 - AcceptedBlocker - Violates the following F20 alpha release criterion for ARM images with a graphical desktop: "A system installed with a release-blocking desktop must boot to a log in screen where it is possible to log in to a working desktop using a user account created during installation or a 'first boot' utility." 16:58:40 ack 16:58:43 ack 16:58:44 ack 16:58:52 #agreed 1002737 - AcceptedBlocker - Violates the following F20 alpha release criterion for ARM images with a graphical desktop: "A system installed with a release-blocking desktop must boot to a log in screen where it is possible to log in to a working desktop using a user account created during installation or a 'first boot' utility." 16:59:07 #topic (985569) libuser wrongly sets "date of last password change" shadow field to 0 on systems without an rtc 16:59:10 #link https://bugzilla.redhat.com/show_bug.cgi?id=985569 16:59:13 #info Proposed Blocker, libuser, ASSIGNED 16:59:40 this bug is on the text version, there is also a patch that has been proposed but not actioned 17:00:06 the consequence is that a user would have to reset their password after initial login, right> 17:00:09 ? 17:00:22 -1/+1 17:00:28 right.. and they dont if they have an rtc/network 17:01:08 -1 blocker, +.5 on FE - depends on how quickly the fix showed up 17:01:21 which ends up being -1/+1, I suppose 17:01:27 we don't have to take the fix 17:01:32 right 17:01:33 sorry, I'll link to criteria in the future on these - this one should likely be FE 17:01:34 other votes? 17:01:53 -1/+1 17:02:06 for the first it's alpha users having to re-set their password is least of their worries ;) 17:02:33 how many arm platforms don't have rtc? 17:02:38 and.. its only if the network was missing, I'd agree 17:02:42 tflink, most 17:02:44 and any of the pre-release stuff are throw away installs anyway 17:03:24 proposed #agreed 985569 - RejectedBlocker AcceptedFreezeException - This doesn't violate any of the F20 alpha release criteria and thus is rejected as a release blocking bug for alpha. However, a tested fix would be considered past freeze. 17:03:28 ack 17:03:36 ack 17:03:37 ack 17:03:44 #agreed 985569 - RejectedBlocker AcceptedFreezeException - This doesn't violate any of the F20 alpha release criteria and thus is rejected as a release blocking bug for alpha. However, a tested fix would be considered past freeze. 17:03:51 ok, that's all of the proposed blockers on my list 17:03:55 on to the proposed FE 17:04:10 oh, there aren't any 17:04:17 QUANTUM FUSE! 17:04:19 on to the accepted blockers, then :) 17:04:38 * tflink skips MODIFIED bugs 17:04:41 tflink, no need to review accepted blockers I think this early in the cycle 17:04:49 * satellit_e https://bugzilla.redhat.com/show_bug.cgi?id=1004421 not yet listed 17:05:03 ( we should do so just before we plan releasing alpha ) 17:05:08 satellit: you proposed it as a beta FE 17:05:14 oops 17:05:50 Viking-Ice: I disagree on accepted blockers, I'd rather catch issues that are stagnating now rather than just before go/no-go 17:05:51 +1 fe 17:06:20 tflink: yep, I tried some decent pinging today on a few 17:06:48 tflink, stagnating might == to people working on the solution 17:06:56 #topic (1004421) f20 SoaS TC3 uses gdm which defaults to gnome. Autologon starts gnome with no chance to choose sugar-desktop 17:06:59 #link https://bugzilla.redhat.com/show_bug.cgi?id=1004421 17:07:02 #info Proposed Freeze Exception, sugar, NEW 17:07:05 +1 FE 17:07:23 would be a blocker for a release-blocking DE, therefore FE for sugar 17:07:43 well that switch is coming a bit late 17:08:14 proposed #agreed 1004421 - AcceptedFreezeException - This prevents SoaS from booting properly and would be a blocker for a release blocking DE, therefore accepted as a FE for F20 alpha. 17:08:19 that runner request in a perfect pony world should have arrived when we branched 17:08:20 ack 17:08:31 ack 17:08:59 ack/nak/patch? 17:09:19 ack 17:09:44 #agreed 1004421 - AcceptedFreezeException - This prevents SoaS from booting properly and would be a blocker for a release blocking DE, therefore accepted as a FE for F20 alpha. 17:10:11 of the accepted blockers, the only one whose status I'm not clear on is 1002055 17:10:36 1000715, 1001081 are waiting on TC3 17:10:49 1002055 is waiting on gnome update 17:10:54 * spstarr_work is going to wait for TC4 unless TC3 fixes anaconda 32bit boot issues or other serious ones im blocked for major testing 17:11:00 1000889 has a patch, should be in next anaconda build 17:11:52 #topic Accepted blockers 17:11:52 I don't see any accepted blockers that need attention right now 17:12:03 #info 1000715, 1001081 are waiting on TC3 17:12:14 #info 1002055 is waiting on gnome update 17:12:26 #info 1000889 has a patch, should be in next anaconda build 17:12:48 it looks like 1000927 is being worked on, nothing needed from us at the moment 17:13:24 btw Anaconda in 20130903 nightly build is broken, after I "Accept" my fate", screen becomes white 17:13:42 in virt-manager 17:13:56 oh, fun - i wonder if TC3 will be DOA 17:13:59 Martix: 32bit? 17:14:02 we'll find out shortly 17:14:04 kparal: 64 17:14:11 hmm 17:14:29 from PXE 17:14:31 any objections to skipping review of accepted blockers for today? 17:14:44 lol 17:14:48 Martix: that kinda sounds like an issue spstarr_work was chasing down? 17:14:48 nope 17:14:50 I come in to see y'all wrapping up 17:14:51 tflink: no 17:15:16 #agreed no more review of accepted blockers is needed today, all seem to be progressing well 17:15:22 #topic Open Floor 17:15:36 Any other topics which should be brought up now? 17:15:37 < tflink> 1002055 is waiting on gnome update 17:15:48 kalev: ? 17:15:54 this is actually done and made it to F20 last night, right before the freeze 17:16:14 tflink: do you have link in bugzilla? 17:16:18 kalev: i thought those builds were being done in a side tag? 17:16:21 unfortunately, TC3 is based on yesterday's F20 compose, so it would be great if you could request one more TC 17:16:34 tflink: yeah, but we finished up before the freeze and merged them back in. 17:16:39 sweetness 17:16:56 yeah, I expect to request a new TC soon 17:17:02 cool 17:17:08 thanks for the update 17:17:18 new tc *ought* to boot on Beagle Bone Black 17:17:29 handsome_pirate: post TC3? 17:17:30 handsome_pirate: great news 17:17:52 we should probably build a smoke livecd to test the x32 gnome issue, though 17:18:02 since it'll be a day or so before TC4 is requested/built 17:18:15 #info F20 Alpha TC3 should be landing shortly, will need testing 17:18:56 #info next blocker review meeting will be 2013-09-11 @ 16:00 UTC unless another meeting is needed on 2013-09-09 17:19:08 tflink: Aye, post TC3 17:19:20 tflink: It depends on if the latest kernel build gets in 17:19:25 if there's nothing else, I'm setting the non-deterministic fuse for [0,5] minutes 17:19:35 3.11.0-3 17:19:38 handsome_pirate: dgilmore pulled in a new kernel update to TC3 17:19:51 Okay, that should have it 17:19:55 you guys are just trying to make me look bad for raising an issue with fesco :) 17:20:08 kparal: tflink also 20130831 build hits white screen, fyi TC2 doesn't 17:20:20 * Viking-Ice turns on that quantum fuse that changes non-deterministic fuse to be even more non-deterministic 17:20:41 I'm affraid, TC3 will have some issue 17:20:45 *same 17:20:59 Martix: yeah, we'll see if TC3 is also affected but I suspect you're right 17:21:33 anyhow, we have multiple fuses which could go off any time :) 17:21:38 thanks for coming everyone! 17:21:43 * tflink will send out minutes shortly 17:21:46 #endmeeting