15:00:29 #startmeeting Fedora QA meeting 15:00:29 Meeting started Mon Aug 26 15:00:29 2013 UTC. The chair is handsome_pirate. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:29 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:39 tflink: Error: Can't start another meeting, one is in progress. 15:00:41 #chair kparal jskladan tflink 15:00:41 Current chairs: handsome_pirate jskladan kparal tflink 15:00:56 ha, 15:00:59 * jskladan already has a chair though ;) 15:01:07 #meetingname Fedora QA Meeting 15:01:07 The meeting name has been set to 'fedora_qa_meeting' 15:01:21 #topic Roll Call 15:01:29 * kparal is here 15:01:32 #info agenda is in wiki 15:01:35 * jskladan "yo dawg, we heard you like chairs..." 15:01:44 #link https://fedoraproject.org/wiki/QA/Meetings/20130826 15:01:58 * spoore is here...listening 15:02:00 * mkrizek is here 15:02:08 * handsome_pirate builds kernels 15:02:23 jskladan: one word - badges! 15:02:49 so ... who's running this show? 15:03:03 tflink: I can, if you want 15:03:09 handsome_pirate: knock yourself out 15:03:13 roger 15:03:19 * pwhalen is here 15:03:29 #topic Previous Meeting Followup 15:03:44 nirik: Ping 15:03:53 The first two items are from adamw 15:04:04 #info adamw to draft the final Cloud changes today 15:04:07 yes? 15:04:16 #info adamw and pschindl to work with anaconda team on anaconda showstoppers 15:04:28 nirik: The third followup from last week is yours 15:04:37 #info nirik or adamw to fix spin-kickstarts user creation bug 15:04:45 adamw did that. 15:04:50 Ah, okay 15:04:58 should be fixed. :) 15:05:04 the showstoppers that info refers to should be fixed 15:05:10 * oddshocks is lurking 15:05:15 And, I'm guessing he took care of the other two since we have a TC 15:05:44 Anything else y'all want to expand on there? 15:05:48 dshea figured out what was killing installations 15:06:04 something about the order of commands to a lightbox window 15:06:30 Ah, cool 15:06:54 Okay, then 15:07:05 #topic Fedora 20 Planning 15:07:15 #info TC1 has landed 15:07:54 The issues so far seem to be the usual crop of TC1 blockers, but also spin problems 15:07:58 validation treadmill time! 15:08:01 For instance, KDE is too large 15:08:08 tflink: +1 15:08:25 the blocker list doesn't look too bad so far 15:08:32 * pwhalen is testing arm images, pxe installs today 15:08:57 * tflink doesn't understand why we're waiting for freeze to add f20 to bodhi, but that's a different issue :) 15:09:11 pwhalen: is there a beaglebone image? 15:09:19 * tflink thought someone was working on that 15:09:22 tflink: Not yet 15:09:37 tflink, we still need a booting kernel for the bb sadly 15:09:44 tflink: I'll bug Kyle about the status of the kernel 15:09:45 that was my question. we received beaglebone blacks, but are they even supported? 15:09:57 kparal: Not yet :( 15:10:11 but they will be 15:10:23 well, it's going to be hard to do QA until then 15:10:34 since we have no other arm devices 15:10:37 * pschindl is here 15:10:38 kparal, agreed. you can also use qemu 15:10:43 and I just got my new SD cards :) 15:10:52 but they can work on my panda as well 15:11:03 Aye 15:11:03 tflink, when panda's fixed 15:11:05 pwhalen: but some qemu arm modules are not packaged in fedora, are they? 15:11:09 and its getting closer 15:11:14 We support vxpress (qemu) 15:11:28 pwhalen: so what's working right now? 15:11:44 all images work with qemu, you need to extract the initramfs and vmlinuz 15:11:54 it sounds like trimslice, panda and bb are all broken ATM 15:12:04 Indeed 15:12:08 working right.. qemu, highbank, trimslice boot but no network 15:12:35 sorry, trimslice kernel is fine, but f20 image wont work til the glibc issue is fixed 15:13:00 so we're 1 week away from freeze and almost none of the supported platforms boot? 15:13:14 * tflink has a bad feeling about this 15:13:15 Indeed 15:13:23 kyle has been working hard on omap, and has now discovered the kernel issue appears to be uboot 15:13:54 related question - why do we have "Trimslice, Panda, Panda ES" in the installation matrix? why only these 3, and why not just a single "ARM" column? https://fedoraproject.org/wiki/Test_Results:Fedora_20_Alpha_TC1_Install?rd=Test_Results:Current_Installation_Test#ARM_disk_images 15:14:16 #info non working ARPM platforms as of today: pandaboard, beaglebone black, trimslice 15:14:28 kparal: different boot methods 15:14:45 and different chips 15:14:54 #undo 15:14:54 Removing item from minutes: 15:15:12 #info non working ARM platforms as of today: pandaboard, beaglebone black, trimslice 15:15:16 panda and panda es could share the same column, and perhaps the name changed to reflect the actual difference, ext vs vfat partitioning for boot 15:15:44 tflink: I see only two different images - vfat a non-vfat 15:15:44 http://dl.fedoraproject.org/pub/alt/stage/20-Alpha-TC1/Images/armhfp/ 15:16:04 pwhalen: in that case it should, otherwise it's very confusing for someone who doesn't know the differences 15:16:06 #info work on omap issue has been progressing, new issues in uboot are causing problems on panda 15:16:23 kparal, agreed 15:16:28 So, we might want to propose that if it isn't working by freeze, we don't support it 15:16:30 ok, I'll adjust the template 15:16:44 kparal, thanks 15:17:14 other arm issues other than non-functional platforms? 15:17:49 tflink, will know more today, really only know the state of the kernel 15:18:20 I have been doing basic boot testing of the rawhide images.. on qemu only as the hw has issues (trimslice glibc issue, panda non booting kernel) 15:18:26 ok, I suspect that x86 is in a similar position 15:18:46 is there any ETA on getting any of that fixed? 15:19:05 not as of yet 15:19:19 * tflink is not thrilled about releasing alpha with a PA not supporting many of its platforms 15:19:28 thrilled at the thought 15:19:52 tflink: Neither am I 15:20:30 things tend to shape up quickly near the end, so much change in the kernel for arm 15:20:43 pwhalen: is there a blogpost somewhere how to run arm images in virt-manager? I'd link that to the test cases 15:20:53 pwhalen: that doesn't really comfort me much 15:21:02 kparal: virt-manager doesn't support arm yet 15:21:11 kparal, I need to write something for f20, you can use the f19 install page 15:21:16 kparal: According to the virt-manager folks, it's coming 15:21:29 kparal: Right now, you have to do qemu manually 15:21:46 kparal, https://fedoraproject.org/wiki/Architectures/ARM/F19/Installation#For_Versatile_Express_Emulation_with_QEMU 15:21:55 * kparal will have a look 15:22:28 #info for more on booting ARM images with qemu: 15:22:35 #link https://fedoraproject.org/wiki/Architectures/ARM/F19/Installation#For_Versatile_Express_Emulation_with_QEMU 15:22:51 is there anything else on ARM that we need to cover? 15:23:34 tflink, I dont think so, I will find out more on the status today 15:23:48 handsome_pirate: shall we move on to cloud and x86? 15:23:53 Sure 15:24:14 has anyone tried the cloud image yet? 15:24:20 tflink: I have not 15:24:31 * tflink assumes not since TC1 wasn't released all that long ago 15:24:37 tflink: Did it ever get sorted out where we can run the cloud images? 15:24:47 * tflink hopes we'll see more participation from the cloud folks than in previous releases 15:24:54 handsome_pirate: I don't believe so, no 15:25:02 Cloud QE offered some help 15:25:20 but the discussion trailed off 15:25:24 well for ARM - in case it would not be ready for Alpha in time - where do we want to go to FESCo before blocking release? (as for ARM contingency plan)? 15:25:59 jreznik: I'd say lets wait for freeze 15:26:21 if we still don't have any working platforms outside qemu and highbank at that point, alert FESCo 15:26:28 Aye, see how things stand next week 15:26:44 did the cloud changes get merged into release criteria? 15:26:57 tflink: agree 15:26:59 * tflink isn't 100% sure what those changes are, just sees item from last week 15:27:21 tflink: https://fedoraproject.org/w/index.php?title=Fedora_20_Alpha_Release_Criteria&diff=349894&oldid=348391 15:27:24 nirik: They did 15:27:36 "Release-blocking cloud images must boot in the [[Infrastructure/FedoraCloud|Fedora OpenStack Cloud]] and in [https://aws.amazon.com/ec2/ Amazon EC2]." 15:27:51 * nirik looks up, hum? 15:27:54 "* Release-blocking cloud images must allow login with the user authentication configuration requested during instance creation." 15:28:18 nirik: handsome_pirate wanted to reply to tflink, I guess 15:28:23 nirik: I think he meant that for me 15:28:50 ah, ok. ;) 15:28:59 Aye 15:29:11 * nirik is easily confused in the morning. ;) 15:29:27 * handsome_pirate doesn't get how he managed nirik from tflink 15:29:33 * handsome_pirate blames tab-completion 15:29:36 we need to start figuring out who's working on the cloud image validation, though 15:29:51 * nirik loads the tc1 images into our cloud. 15:29:53 handsome_pirate: n and t are on different parts of the keyboard, though :) 15:30:10 tflink: If I can get somewhere to do it, I'll test cloud 15:30:12 tflink perhaps he's using a modified dvorak layout 15:30:24 roshi: Nope :) 15:30:45 do we know if the AMIs were uploaded to amazon? 15:30:50 well, then you have no excuse handsome_pirate 15:30:50 * tflink assumes not 15:31:16 tflink: dgilmore says that there are compose issues with cloud images 15:31:24 tflink: He says composition is broken 15:31:34 hum, no 32bit cloud image. ;( 15:32:48 any volunteers to coordinate with the cloud folks for testing images and coordinating which images are supposed to be where? 15:32:56 * handsome_pirate waves 15:33:27 #action handsome_pirate to coordinate with cloud folks for cloud testing and where images are to be located 15:33:29 #action handsome_pirate to coordinate with releng and cloud folks to figure out who's doing testing and which images are supposed to be where 15:33:32 damnation 15:33:36 #undo 15:33:36 Removing item from minutes: 15:33:51 anything else on cloud? 15:34:06 * tflink assumes not 15:34:10 how about x86? 15:34:20 * tflink just started downloading images 15:34:29 any obvious showstoppers so far? 15:34:38 I've heard mutterings that it's still not installible 15:34:48 fun 15:34:54 for some reason, it failed to start over PXE on two different machines 15:34:59 VM worked well 15:35:06 * kparal still investigating 15:35:28 kparal: cool, sounds like some progress at least :) 15:36:02 tflink: < rrknight3> I'm completely unable to install TC1. DeviceFormat has no attribute mount in both graphics and text mode on x86_64. 15:36:03 anything else for x86 testing/planning? 15:36:22 custom partitioning? 15:36:23 (that's from this morning in #fedora-qa 15:36:26 ) 15:36:35 He didn't say 15:37:46 I guess we'll find out as more people test 15:38:01 it's not like TC1 has been out for long 15:38:06 Aye 15:38:16 any other issues to cover for F20? 15:39:01 * handsome_pirate thinks it's time for: 15:39:04 handsome_pirate: open floor? 15:39:06 #topic Open Floor 15:39:23 Anyone have anything? 15:39:56 test 15:40:04 * handsome_pirate sets the magical unicorn rainbow sparkle timer for five minutes 15:40:05 test day request should be going out soon 15:40:18 roshi: ^^ correct? 15:40:26 tflink did you get me added to the group? 15:40:31 yep 15:40:32 * roshi is at work and can't check atm 15:40:53 Group? 15:40:55 yeah - I'll get that out when I get home 15:41:03 handsome_pirate: for modifying the qa calendar 15:41:11 Ah 15:41:22 tflink: Am I part of that group? 15:41:40 * handsome_pirate is in several qa groups and can't remember which does what 15:41:52 handsome_pirate: I don't think so 15:42:15 tflink I'll send you a draft when I get home - since it'll be my first 15:42:28 roshi: sounds like a plan 15:42:41 roshi: I did it a few releases ago; just make sure you keep up on the details 15:43:19 handsome_pirate sounds good 15:43:23 * handsome_pirate once again sets the magic pony fuse 15:43:31 * roshi will get the hang of it no problem 15:44:36 Anything else? 15:44:49 Going once 15:44:54 Twice 15:44:58 #endmeeting