QA/Meetings/20090722

From FedoraProject

Jump to: navigation, search

Contents

Attendees

  • Adam Williamson (adamw)
  • David Pravec (dpravec)
  • Jesse Keating (f13)
  • jeff_hann
  • James Laska (jlaska)
  • Josh Boyer (jwb)
  • Mel Chua (mchua)
  • Kevin Fenzi (nirik)
  • John Poelstra (poelcat)
  • Sebastian Dziallas (sdziallas)
  • Jóhann Guðmundsson (viking-ice)
  • Will Woods (wwoods)

Regrets:

  • Kamil Paral (kparal)


Agenda

Proposed meeting agenda

Previous meeting follow-up

  1. [jlaska] - Notify team list (fedora-test-list) to solicit volunteers for help with Debugging pages (see http://tinyurl.com/n6n24f)

adamw suggested what would be tricky about this is finding out what the necessary info _is_ in the cases where we want debugging pages. adamw and viking_ice indicated their movie preference (see below). f13 suggested the idea of conducting maintainer interviews to use as information gathering sessions to collect debugging information.

  1. [jlaska] - Create a QA:Meeting matrix similar to Bugzappers_meeting_matrix to determine best fit for meeting slots

adamw suggested a nice web app for coordinating meeting times in the future - http://whenisgood.net

QA Meeting time change

See fedora-test-list mail and QA_Meeting_Matrix

The team selected Monday at 16:00 UTC as the new QA meeting time.

In the news

Alpha blocker bug day#2

In preparation for this Friday's blocker bug day#2, jlaska asked the group for volunteers to help announce and summarize the event (in accordance with the draft SOP [1]).

  • Adamw offered to send out bug zapper announcements
  • Jlaska will send out a summary of the blocker bug meeting
  • Poelcat noted the meeting time has changed to 11:00am (EDT) to avoid a conflict with the FESCO meeting

F13 suggested it would be good to perform some rawhide testing prior to. Jlaska reminded folks about Liam's email requesting installer testing for the Alpha Candidate (see https://www.redhat.com/archives/fedora-test-list/2009-July/msg00429.html). Some discussion followed about a Alpha blocker RHBZ #513020

Alpha test compose

The F12 Alpha schedule has a test compose being handed off next week (2009-07-29). See Liam's announcement at https://www.redhat.com/archives/fedora-test-list/2009-July/msg00429.html inviting testers to contribute installation test results against the upcoming Alpha Candidate.

F13 had issue with the announcement:

  1. [Liam] called it a candidate.
  2. [Liam]'s inviting the world to participate, but we're not delivering it to the world. It's not being mirrored.

F13 indicated that download servers fail to keep up with the demand, and wwoods suggested the slower download speed prevents testers from testing the ISO's. Some discussion followed, with several participants offering suggestions (mirrors or bittorrent) and others offering historical perspective to the problem. Poelcat and jlaska suggested this is more an issue for the fedora-infrastructure team to devise a clever solution for ... and not for the QA or RelEng teams to solve.

Jlaska took an action item to engage with fedora-infrastructure team for a solution.

Adamw asked why we are asking for install testing if blocked by RHBZ #513020. Jlaska offered that Liam is getting the pages and test coordination started early

Jlaska suggested point#1 is a minor wording oversight, and point#2 is not a concern for QA.

Dracut gone raw

Features/Dracut is planned to hit rawhide in the next few days. Viking_ice requested discussion around dracut in the meeting.

According to hhoyer, dracut will be landing in rawhide any day now. hhoyer also addressed some of my comments in the feature page (https://fedoraproject.org/wiki/Talk:Features/Dracut) and indicated additional information on a contingency plan and test plan would be arriving soon. Harald also offered his thoughts on the current state of dracut tests:

Thorny areas (potentially buggy):

  • strange raid setups
  • incomplete/degraded raid setups

Should work:

  • MD, DM Raid
  • LVM
  • cryptoLUKS
  • NFS
  • iSCSI
  • NBD

untested:

  • everything else.

Not yet implemented:

  • loop mount
  • FCoE
  • ipv6

Viking_ice asked additional questions around testing dracut, including:

  1. Are we rolling live iso' s with dracut (including gnome, kde and XFCE spins)?
  2. if they are built using our build system?

Viking_ice volunteered to reach out ot Harald for additional information to help guide Dracut testing now and for the Dracut test day.

AutoQA update

Updates from wwoods on initial automation efforts Updates from dpravec on autotest package testing

dpravec suggested many co-workers in Brno feel we should use beaker and not autotest. Jlaska noted that the beaker project is still maturing and does not yet meet the full test automation needs of the israwhidebroken.com project. Wwoods later noted that beaker is inventory / system provisioning, it currently lacks a test harness and scheduler.

dpravec noted there was no updates on the testing of the autotest package, and would have more results by Monday.

Wwoods updated the group that the first 4 test cases of QA:Rawhide_Acceptance_Test_Plan have been automated and are running in demo instances of the autotest server. The next step will be automating the installation test cases using some incantation of libvirt/virt-manager.

F13 expressed concern we're falling short of "do something useful ASAP". Jlaska indicated that the group is progressing towards the israwhidebroken.com target date of Aug 18th. F13 stated ... this whole thing started back at FUDCon Boston with the idea of doing something useful in 3 weeks, useful in 3 months, useful in 3 quarters. So I feel that I've failed miserably in delivering something useful to our maintainers based on those timelines. Jlaska, wwoods and adamw pointed out some project highlights since FAD and noted things are progressing towards stated objectives. F13 concluded expressing frustrations that his window to help the project is closing as the release engineering commitments increase.

Open discussion

Sugar on a stick Test Day

https://www.redhat.com/archives/fedora-test-list/2009-July/msg00443.html

sdziallas informed the team he has been doing the Sugar on a Stick work for Sugar Labs lately, which is basically a rebranded Fedora base system with the Sugar packages on top. We're really interested in maintaining a good connection to the Fedora community (given that I started as a Fedora guy, too)... and would also like to test the Sugar implementation in Fedora.

Discussion followed on how best to offer a test day slot, the result is a Sept 3, 2009 Sugar on a stick test day slot.

Mchua noted the event could be used to test the state of mediawiki+semantic integration.

Improving Rawhide

dpravec noted that from talking with Fedora users, seems to me that best what we can do to improve quality of releases is to make some smoother rawhide repository, which will contain only packages without serious bug or dependency problems. similar to debian unstable/testing... this would enourage people to test latest (or week old) packages.

f13 noted we're going to get essentially that if his no frozen rawhide proposal goes though .

Discussion followed comparing difference between the debian and fedora development models.

<Your topic here>

Upcoming QA events

Action items

  • [Viking-Ice] - updating meeting channel page w/ new QA time
  • [jlaska] - discuss ideas with fedora-infrastructure for getting test candidates out to testers

IRC Transcript

jlaska | #startmeeting Jul 22 12:01
zodbot | Meeting started Wed Jul 22 16:01:10 2009 UTC. The chair is jlaska. Information about MeetBot at http://wiki.debian.org/MeetBot. Jul 22 12:01
zodbot | Useful Commands: #action #agreed #halp #info #idea #link #topic. Jul 22 12:01
jlaska | #meetinggtopic Fedora Quality Assurance Jul 22 12:01
adamw | mor-diddly-orning Jul 22 12:01
* | f13 Jul 22 12:01
f13 | jlaska: #topic Jul 22 12:01
jlaska | #topic Gathering able bodied testers Jul 22 12:01
--- | zodbot has changed the topic to: Gathering able bodied testers Jul 22 12:01
f13 | hrm. Jul 22 12:01
* | Viking-Ice half in half in cloud.. Jul 22 12:01
f13 | jlaska: might want to end the meeting real quick Jul 22 12:01
jlaska | #meetingtopic Fedora QA Meeting Jul 22 12:01
--- | zodbot has changed the topic to: Gathering able bodied testers (Meeting topic: Fedora QA Meeting) Jul 22 12:01
f13 | your #startmeeting needs... there you go Jul 22 12:01
* | poelcat Jul 22 12:01
jlaska | apparently the meeting bot can't fix my poor spelling Jul 22 12:02
* | f13 wonders if there are two meetings going now... Jul 22 12:02
jlaska | odd Jul 22 12:02
jlaska | #endmeeting Jul 22 12:02
--- | zodbot has changed the topic to: Channel is used by various Fedora groups and committees for their regular meetings | Note that meetings often get logged | For questions about using Fedora please ask in #fedora | See http://fedoraproject.org/wiki/Meeting_channel for meeting schedule Jul 22 12:02
zodbot | Meeting ended Wed Jul 22 16:02:23 2009 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . Jul 22 12:02
zodbot | Minutes: http://meetbot.fedoraproject.org/fedora-meeting/2009-07-22/fedora-meeting.2009-07-22-16.01.html Jul 22 12:02
zodbot | Minutes (text): http://meetbot.fedoraproject.org/fedora-meeting/2009-07-22/fedora-meeting.2009-07-22-16.01.txt Jul 22 12:02
jlaska | #startmeeting Jul 22 12:02
zodbot | Log: http://meetbot.fedoraproject.org/fedora-meeting/2009-07-22/fedora-meeting.2009-07-22-16.01.log.html Jul 22 12:02
zodbot | Meeting started Wed Jul 22 16:02:26 2009 UTC. The chair is jlaska. Information about MeetBot at http://wiki.debian.org/MeetBot. Jul 22 12:02
zodbot | Useful Commands: #action #agreed #halp #info #idea #link #topic. Jul 22 12:02
f13 | jlaska: you're doing it wrong Jul 22 12:02
jlaska | story of my life Jesse Jul 22 12:02
f13 | jlaska: your #startmeeting command needs the subject as an argument Jul 22 12:02
jlaska | it doesn't matter Jul 22 12:02
adamw | as the vicar said to the nun Jul 22 12:02
f13 | otherwise the /topic will look all weird Jul 22 12:02
jlaska | #meetingtopic Fedora QA meeting Jul 22 12:03
--- | zodbot has changed the topic to: (Meeting topic: Fedora QA meeting) Jul 22 12:03
Viking-Ice | not the nun said to vicar Jul 22 12:03
jlaska | #topic Gathering people Jul 22 12:03
--- | zodbot has changed the topic to: Gathering people (Meeting topic: Fedora QA meeting) Jul 22 12:03
f13 | ah, guess that will work too Jul 22 12:03
jlaska | okie dokie ... sorry for the noise Jul 22 12:03
* | f13 Jul 22 12:03
jlaska | okay, so far I've got adamw, f13, poelcat, Viking-Ice, $me Jul 22 12:03
jlaska | I should say we have .5 Viking-Ice Jul 22 12:04
* | sdziallas waves to jlaska for sugar related stuff (ping me if needed) Jul 22 12:04
jlaska | sdziallas: welcome! I've added your open discussion topic Jul 22 12:04
sdziallas | jlaska: awesome, thanks! :) Jul 22 12:05
jlaska | I suspect we have wwoods lurking ... and I don't see dpravec just yet ... hopefully he can join later Jul 22 12:06
jlaska | #topic Previous meeting follow-up Jul 22 12:06
--- | zodbot has changed the topic to: Previous meeting follow-up (Meeting topic: Fedora QA meeting) Jul 22 12:06
jlaska | I feel like there was more to capture from last week ... but my notes scrubbing only has 2 items Jul 22 12:06
jlaska | so feel free to chyme in with any updates Jul 22 12:06
* | jlaska working off of proposed agenda - https://www.redhat.com/archives/fedora-test-list/2009-July/msg00438.html Jul 22 12:07
jlaska | a few quick ones ... Jul 22 12:07
jlaska | # [jlaska] - Notify team list (fedora-test-list) to solicit volunteers for help with Debugging pages Jul 22 12:07
jlaska | posted something in my blog about this last week ... inviting contributions Jul 22 12:08
jlaska | http://jlaska.livejournal.com/5693.html Jul 22 12:08
adamw | what might be a bit tricky about this is finding out what the necessary info _is_ in the cases where we want debugging pages Jul 22 12:08
jlaska | adamw: yeah, I was wondering if a template might help? Jul 22 12:08
adamw | i was looking at writing one or two yesterday, but i don't actually know what the required info is, for the three desired pages Jul 22 12:08
adamw | no, that's not the problem, laying out the page is easy Jul 22 12:08
adamw | the question is what actual information is needed for, say, a bug in rpm Jul 22 12:09
Viking-Ice | That's not the test list :) and no pr0n on your blog to attract crowd... Jul 22 12:09
adamw | we probably need to co-ordinate with maintainers on that, unless we have someone who knows Jul 22 12:09
adamw | Viking-Ice: what, you haven't been following jlaska's adult entertainment career? Jul 22 12:09
jlaska | adamw: we're not going there :) Jul 22 12:09
adamw | Real Men Do It In Red Hats is a modern classic Jul 22 12:09
jlaska | LOL! Jul 22 12:10
adamw | yay, i scared him off =) Jul 22 12:10
f13 | adamw: when you picture that with the Red Hat Society, that's very very wrong Jul 22 12:10
* | jlaska purges Jul 22 12:10
adamw | hehe Jul 22 12:10
jlaska | adamw: do you have any thoughts on connecting with maintainers ... Jul 22 12:10
adamw | send 'em an email? :) Jul 22 12:10
jlaska | is that something to add as guidlines for 'how to get started' kind of deal? Jul 22 12:10
jlaska | yeah Jul 22 12:10
jlaska | to me, it seems like a good activity awaiting contributors who already have skills in a particular area Jul 22 12:11
jlaska | but you are right ... the point of hte exercise is to make processing bugs easier for maintainers Jul 22 12:12
jlaska | I've had no takers on the volunteer offer yet ... but will certainly point folks in the right direction :) Jul 22 12:12
jlaska | okay next up ... (another quickie) Jul 22 12:12
jlaska | # [jlaska] - Create a QA:Meeting matrix similar to Bugzappers_meeting_matrix to determine best fit for meeting slots Jul 22 12:12
jlaska | done ... See https://fedoraproject.org/wiki/QA_Meeting_Matrix Jul 22 12:13
jlaska | and thanks folks for adding your availability to the matrix Jul 22 12:13
Viking-Ice | as said before existing maintainers can be pinged to provide this info and new components and their maintainers being introduced into fedora should provide this info Jul 22 12:13
adamw | if this ever comes up again, btw, i came across an awesome system someone on the docs team used Jul 22 12:13
jlaska | Viking-Ice: right on ... I agree Jul 22 12:13
adamw | http://whenisgood.net Jul 22 12:13
jlaska | cute! Jul 22 12:14
adamw | it sets up the matrix for you, and automatically gives everyone who uses it their local time Jul 22 12:14
jlaska | seems to support more TZ's than timeanddate.com Jul 22 12:14
f13 | I think the best way to do it is to invite the maintainer/developer to do an interview style session, where we get them to just brain dump about things and then we pour over it later to get it into the wiki pages. Jul 22 12:14
Viking-Ice | honestly not seeing how this solves anything Jul 22 12:14
f13 | asking them to fill in the wiki pages will not be as successful. Jul 22 12:15
f13 | oh sorry, new topic. Jul 22 12:15
jlaska | f13: that's not a bad idea tbh Jul 22 12:15
jlaska | I was thinking about it as an activity that any QA contributors could engage in if looking to "get involved" Jul 22 12:15
jlaska | hopefully that's not wishful thinking, I like having some things out there that are worthwhile and don't require a huge investment Jul 22 12:16
jlaska | #idea Host developer/maintainer interviews to identify what information is required for good bug reports Jul 22 12:17
jlaska | okay ... any other items from last weeks follow-up that weren't covered yet? Jul 22 12:17
jlaska | okay ... let's dive in ... Jul 22 12:18
jlaska | #topic QA Meeting time change Jul 22 12:18
--- | zodbot has changed the topic to: QA Meeting time change (Meeting topic: Fedora QA meeting) Jul 22 12:18
Viking-Ice | Monday's at 16:00 UTC Jul 22 12:18
jlaska | thanks to everyone who updated availability in https://fedoraproject.org/wiki/QA_Meeting_Matrix Jul 22 12:18
jlaska | the top candidates seem to be what Viking-Ice noted ... and Jul 22 12:19
jlaska | * Mon 16:00 UTC (12 PM EDT) Jul 22 12:19
jlaska | * Thu 15:00 UTC (11 AM EDT) Jul 22 12:19
jlaska | * Fri 15:00 UTC (11 AM EDT) Jul 22 12:19
f13 | I'd prefer the first Jul 22 12:20
f13 | the first will cut into my cycling time for part of the year, the others will cut into my sleep time Jul 22 12:20
adamw | yeah, i like the first, but i can live with any of them Jul 22 12:20
jlaska | I was secretly hoping to get out of hte 16:00 UTC hour Jul 22 12:20
jlaska | but can do all of the above Jul 22 12:21
Viking-Ice | I prefer monday's as well + https://fedoraproject.org/wiki/Meeting_channel either has been changed or we did mistakes there's nothing schedualed at 17:00 which allows for extended of the meetings since we have been slipping over our meeting time from time to time.. Jul 22 12:21
adamw | good point, a spare hour after the meeting is always nice Jul 22 12:22
* | jlaska shudders at the thought Jul 22 12:22
jlaska | alrighty ... Jul 22 12:22
jlaska | #agreed the new meeting time will be on Monday at 16:00 UTC (12 PM EDT) Jul 22 12:22
jlaska | next topic ... Jul 22 12:23
jlaska | just some brief news updates Jul 22 12:23
jlaska | #topic Alpha blocker bug day#2 Jul 22 12:23
--- | zodbot has changed the topic to: Alpha blocker bug day#2 (Meeting topic: Fedora QA meeting) Jul 22 12:23
jlaska | I believe we are scheduled for another blocker bug day this Friday Jul 22 12:23
* | poelcat notes change of time to 11 AM EDT Jul 22 12:23
poelcat | to avoid clash w/ fesco meeting Jul 22 12:23
jlaska | ah good stuff Jul 22 12:23
f13 | ah Jul 22 12:23
f13 | let me update my phone Jul 22 12:24
* | nirik asks someone to update the meeting channel page for the new meeting time. Jul 22 12:24
jlaska | would anyone like to volunteer sending ou the announcement? Jul 22 12:24
jlaska | Viking-Ice: do you mind updating the meeting channel wiki page? Jul 22 12:24
poelcat | nirik: it meets on #fedora-bugzappers... does that still apply? Jul 22 12:24
Viking-Ice | sure Jul 22 12:24
jlaska | Viking-Ice: thanks! Jul 22 12:24
nirik | poelcat: no, I meant the qa meeting time... ;) Jul 22 12:24
jlaska | we've got a few prep tasks for the blocker bug day ... Jul 22 12:25
jlaska | * send out announcements Jul 22 12:25
jlaska | * take notes and send summary Jul 22 12:26
jlaska | Jul 22 12:26
jlaska | anyone interested in either? Jul 22 12:26
jlaska | btw ... thanks to adamw for providing a summary from last week (https://www.redhat.com/archives/fedora-test-list/2009-July/msg00347.html) Jul 22 12:26
f13 | sorry I can't take on any of that :/ Jul 22 12:26
adamw | I can do them if necessary Jul 22 12:26
Viking-Ice | doen Jul 22 12:26
Viking-Ice | mean done Jul 22 12:27
* | jlaska would like to give others the opportunity Jul 22 12:27
jlaska | Viking-Ice: thx :) Jul 22 12:27
jlaska | adamw: you want to swap duties this week? Jul 22 12:27
jlaska | you take the announcement and I'll do the summary? Jul 22 12:27
adamw | sure Jul 22 12:27
adamw | any combination is fine for me Jul 22 12:27
jlaska | cool, thanks Jul 22 12:28
jlaska | any other things we should be thinking about for blocker bug day#2? Jul 22 12:28
f13 | would be good to have some rawhide testing done by then Jul 22 12:28
f13 | of course I think rawhide is in cannot install, full stop mode right now :/ Jul 22 12:29
jlaska | f13: yeah ... I've kicked the tires on that a little and asked lili to bump up the install testing in response to your requests Jul 22 12:29
jlaska | we've got a few blocker bugs on the list now to work with Jul 22 12:29
jlaska | and you and clumens seem to be working the networking issue now Jul 22 12:29
jlaska | good stuff! Jul 22 12:29
f13 | sortof. Jul 22 12:30
f13 | it's a head scratcher Jul 22 12:30
adamw | which issue is this? Jul 22 12:30
jlaska | f13: yeah it had be lost ... but that doesn't take much ;) Jul 22 12:30
jlaska | 513020 Jul 22 12:30
adamw | ah Jul 22 12:30
adamw | jlaska: i am impressed by your painstaking reproduction of the ascii failure dialog =) Jul 22 12:31
jlaska | cut'n'paste my friend Jul 22 12:31
jlaska | thank the newt gods Jul 22 12:31
jlaska | dpravec: welcome! Jul 22 12:32
dpravec | hello, i am sorry for being late Jul 22 12:32
jlaska | dpravec: no worries ... we've just agreed earlier to move the meeting time to Monday @ 16:00 UTC Jul 22 12:32
dpravec | hurray Jul 22 12:32
jlaska | okay ... next update ... Jul 22 12:32
jlaska | #topic Alpha test compose Jul 22 12:32
--- | zodbot has changed the topic to: Alpha test compose (Meeting topic: Fedora QA meeting) Jul 22 12:32
jlaska | the F12 Alpha schedule has a test compose being handed off next week (2009-07-29) Jul 22 12:33
jlaska | Liam started the ball rolling on coordinating install testing ... so if you'd like to pitch in ... give a gander to https://www.redhat.com/archives/fedora-test-list/2009-July/msg00429.html Jul 22 12:33
f13 | I have some issue with that Jul 22 12:34
f13 | 1) he called it a candidate. Jul 22 12:34
f13 | 2) he's inviting the world to participate, but we're not delivering it to the world. It's not being mirrored. Jul 22 12:34
dpravec | and is it installable? Jul 22 12:35
dpravec | i was not the only one to try... and we all failed Jul 22 12:35
Viking-Ice | that's what we are trying to find out := Jul 22 12:35
jlaska | yeah I think #1 is just a minor goof ... we can correct the wording Jul 22 12:35
adamw | yeah, that's what i was thinking...we seem to have a complete blocker bug on installation atm, don't we? so what install testing can we do? Jul 22 12:35
jlaska | adamw: great point ... so if we fail to hit the 07-29 target ... things move Jul 22 12:35
adamw | there _is_ an updates.img for the blocker bug, so we should be testing that, i guess Jul 22 12:36
jlaska | Liam is just getting the pages and coordination started ... so this is good fedback Jul 22 12:36
f13 | right, we have to fix the current issue by then Jul 22 12:36
jlaska | f13: to your second point ... I think that's an area where QA needs a clever solution Jul 22 12:36
f13 | but we're still not going to be able to deliver those trees to very many people. Jul 22 12:36
Viking-Ice | 3) should have provided this link :) https://fedoraproject.org/wiki/Anaconda/BugReporting Jul 22 12:36
Viking-Ice | as well Jul 22 12:37
jlaska | Viking-Ice: feel free to add that link to Liam's announcement Jul 22 12:37
dpravec | btw few hours ago i filled my disk, getting 'rpmdb open failed' later ->which is really 'fun'. luckily i solved this one, but this should not happen... Jul 22 12:38
jlaska | f13: do you have concerns that the server hosting the ISO images won't be able to hold up to testers downloading the images? Jul 22 12:38
f13 | yes Jul 22 12:38
f13 | jlaska: it happens every time the location leaks when I'm trying to deliver stuff to QA Jul 22 12:39
f13 | via Fedora infrastructure Jul 22 12:39
dpravec | so should we use bittorrent? Jul 22 12:39
wwoods | bittorrent doesn't work for this. Jul 22 12:39
jlaska | f13: is there something we can reach out to for help from the infrastructure guys? Jul 22 12:39
Viking-Ice | more bandwith.. Jul 22 12:39
Viking-Ice | the iso's are not the only problem.. Jul 22 12:39
f13 | jlaska: theyr'e going to tell us to use the mirror system Jul 22 12:39
dpravec | why is bittorrent not the way to go? Jul 22 12:39
f13 | which we don't really have time to do Jul 22 12:39
jlaska | so let's use the mirror system? Jul 22 12:39
wwoods | we tried before - demand isn't high enough, so there aren't enough peers/seends Jul 22 12:39
Viking-Ice | testerst should all be using the same bits for testing instead of relying on outdated mirrors.. Jul 22 12:39
f13 | dpravec: lots of little files, very few seeders Jul 22 12:39
wwoods | so nobody ever manages to finish the image before it's obsoleted. Jul 22 12:40
f13 | jlaska: we don't have the 3 or 4 days it would take for the mirrors to sync up Jul 22 12:40
jlaska | okay, so we have ISO's ... and people want to download and test them Jul 22 12:40
dpravec | ah, ic Jul 22 12:40
jlaska | but we can't keep up with that demand? Jul 22 12:40
f13 | this was supposed to be a compose delivered to a few high bandwidth testers Jul 22 12:40
* | poelcat thinks this issue should be opened in a ticket with infrastructure... let the experts handle this problem Jul 22 12:40
f13 | jlaska: the problem is you get a lot of pile on downloaders who provide no actual feedback Jul 22 12:40
f13 | they just jump on teh bits because it's there, and use up all the resources Jul 22 12:41
dpravec | hmm arent those 2 problems? i would use bittorrent for isos Jul 22 12:41
wwoods | right, the test composes aren't intended for a worldwide audience, only people with enough bandwidth to get the images quickly Jul 22 12:41
jlaska | f13: sure ... that's a valid concern .. and something QA should be thinking about right? Jul 22 12:41
wwoods | and enough brains to know what they're doing Jul 22 12:41
jlaska | the test composes are for testers Jul 22 12:41
poelcat | we've wasted days arguing about this here and on the mailing lists Jul 22 12:41
dpravec | you can have one 'seeder' on server... Jul 22 12:41
f13 | I guess a simple solution is to password protect the directory and rsync Jul 22 12:41
f13 | or FAS protect it or something Jul 22 12:41
jlaska | ooh Jul 22 12:41
jlaska | perhaps a tie-in now with the FAS QA groups? Jul 22 12:42
jlaska | that's not a bad idea Jul 22 12:42
wwoods | right, we've talked about that before. private mirror for QA only. Jul 22 12:42
jlaska | cool, yeah let's take this offline and see if we can get something actionable out of this Jul 22 12:42
wwoods | except that reduces to the "what do you have to do to prove yourself worthy of QA team membership" problem Jul 22 12:42
jlaska | since I don't think we'll figure it out in 18 minutes Jul 22 12:42
* | Viking-Ice thinks we are heading down the wrong road restrict access to those img.. Jul 22 12:43
jlaska | it's _a_ solution to a bandwith problem Jul 22 12:43
Viking-Ice | we want to expose testing to as many as possible... Jul 22 12:43
wwoods | we're not restricting access, we're preventing overload Jul 22 12:43
wwoods | a subtle but important difference Jul 22 12:43
jlaska | f13: do you want to reach out to the infrastructure guys ... or should I? Jul 22 12:44
wwoods | you're still welcomed and encouraged to spread those images as far and wide as you want Jul 22 12:44
jlaska | you may have already done this ... so it's possible I'm just late to the party Jul 22 12:44
wwoods | they're for everyone who can get them Jul 22 12:44
* | mchua listens Jul 22 12:44
wwoods | *but*, in order to guarantee that our *testers* can get them, we need to give the testers priority accecss Jul 22 12:44
jlaska | mchua: hey there :) Jul 22 12:44
dpravec | hmm but still, bittorent can work if we will have stable seeders (clients which we will run on servers) Jul 22 12:44
wwoods | otherwise everyone hammers the server and *nobody* gets them Jul 22 12:44
Viking-Ice | wwoods: ofcourse Jul 22 12:44
dpravec | at least for iso images Jul 22 12:44
wwoods | and then, by allowing access to *more* people we end up with *less* testing Jul 22 12:44
wwoods | which is unacceptable. Jul 22 12:45
jlaska | wwoods: all valid points ... I think there is a solution out there Jul 22 12:45
jlaska | perhaps adding value to what it means to be in the QA group is one way Jul 22 12:45
jlaska | okay ... going to table this for now Jul 22 12:45
jlaska | if anyone would like to take an action item to talk to the infrastructure folks to see what we can work out ... let me know Jul 22 12:46
jlaska | otherwise, I'll plan to talk with them Jul 22 12:46
wwoods | dpravec: sure, we could combine both approaches - private mirror for testers, but all testers *must* seed the torrent Jul 22 12:46
--- | stickster is now known as stickster_food Jul 22 12:46
adamw | to be enforced how? Jul 22 12:46
* | jlaska looks for the lid to the can of worms Jul 22 12:47
dpravec | you can cut them from the bittorent server if they do not seed enough Jul 22 12:47
adamw | i don't think anyone wants to have the 'Bittorrent Police' job title Jul 22 12:47
f13 | jlaska: I've got too much on my plate as it is, can you? Jul 22 12:47
jlaska | f13: you got it Jul 22 12:47
jlaska | f13: err ... meaning I got it Jul 22 12:47
jlaska | ;) Jul 22 12:47
adamw | dpravec: see above Jul 22 12:47
adamw | dpravec: who wants to spend their time checking on that? Jul 22 12:47
dpravec | can be automated Jul 22 12:47
adamw | and dealing with the inevitable emotional appeals? Jul 22 12:47
adamw | ^^ Jul 22 12:47
dpravec | there are such servers imho Jul 22 12:48
jlaska | okay gang ... let's continue the bittorrent part when we hear back from infrastructure Jul 22 12:48
wwoods | fine, s/must/are urged to/ Jul 22 12:48
adamw | yeah, it's a tactic commonly used by, ahem, pirate sites =) Jul 22 12:48
jeff_hann | personally I don't think restricting access is a good idea. MHO Jul 22 12:48
dpravec | adamw: not only pirates... Jul 22 12:48
Viking-Ice | people the problem is lack of bandwith so let's fix that problem let's get this iso's on edu bandwith! Jul 22 12:48
wwoods | we tried unrestricted access and nobody could get the bits in time for them to be useful Jul 22 12:48
jlaska | jeff_hann: that's certainly an area we need to be in tune with Jul 22 12:48
wwoods | so maybe think again. Jul 22 12:48
Viking-Ice | selective testers on edu bandwith offer the iso's Jul 22 12:49
poelcat | jlaska: what is the next topic? Jul 22 12:49
dpravec | put the iso on some really accessible mirros? Jul 22 12:49
jlaska | #topic Dracut gone raw Jul 22 12:49
--- | zodbot has changed the topic to: Dracut gone raw (Meeting topic: Fedora QA meeting) Jul 22 12:49
jlaska | Viking-Ice: you had asked for some discussion around the incoming rawhide dracut changes? Jul 22 12:49
Viking-Ice | yes Jul 22 12:49
Viking-Ice | we dont have any implementation plan from the maintainer Jul 22 12:50
Viking-Ice | no goals to see if they are made before the switch Jul 22 12:50
jlaska | I spoke w/ harald this morning about the open questions in the feature page Jul 22 12:50
Viking-Ice | what's needed to be done for a successful switch ( feed back etc ) Jul 22 12:50
jlaska | https://fedoraproject.org/wiki/Talk:Features/Dracut Jul 22 12:50
jlaska | harald correctly pointed out some bozo used the wrong rpm options Jul 22 12:51
jlaska | where bozo == $me Jul 22 12:51
jlaska | so to the first point, the `rpm --whatrequires mkinitrd` list is empty Jul 22 12:51
jlaska | as for a contingency plan ... Harald suggested looking at the Dracut page that Viking-Ice created Jul 22 12:51
jlaska | but I think this could use someone to follow-up with Harald on Jul 22 12:52
jlaska | at least it wasn't clear to me, in the event we need to pull out this change ... what are the high level changes anticipated Jul 22 12:52
jlaska | Viking-Ice: do you know more here? Jul 22 12:52
Viking-Ice | We are still missing some kind of implementation plan that should have milestones which we can then keep track on and test to see if they are met.. Jul 22 12:52
adamw | so, to summarize, you both agree we don't have test details or a contingency plan, and we need to ask the feature owners to deal with this Jul 22 12:53
jlaska | adamw: you got it Jul 22 12:53
jlaska | harald pointed out some areas of testing concern for when this lands in rawhide (real soon now) Jul 22 12:53
jlaska | http://fpaste.org/paste/19555 Jul 22 12:53
jlaska | I suspect these will be strong candidates for the future test day Jul 22 12:54
adamw | and by 'we' i mean 'one of you two' Jul 22 12:54
adamw | :D Jul 22 12:54
jlaska | ha! Jul 22 12:54
jlaska | I'm approaching full ... Viking-Ice do you think you can get some chat with w/ Harald in the next few days? Jul 22 12:55
Viking-Ice | Are we rolling live iso' s with dracut ? Jul 22 12:55
jlaska | for the test day, we plan to Jul 22 12:55
jlaska | but for rawhide ... there are none planned afaik Jul 22 12:55
jlaska | anyone interested in taking an action item to chat w/ harald about Dracut ... which can be used to help outline the dracut test day plan? Jul 22 12:56
Viking-Ice | so sugar kde gnome xfce contain dracut img ? Jul 22 12:56
Viking-Ice | live spins Jul 22 12:56
jlaska | Viking-Ice: that's my understanding ... that new-kernel-pkg would be using dracut (instead of mkinitrd) Jul 22 12:57
Viking-Ice | if they are built using our build system which brings us to another topic Jul 22 12:57
Viking-Ice | I'm sure that those that are "locally" (DE people ) building iso's arent using dracut Jul 22 12:58
jlaska | Viking-Ice: is this something you can help with ... in terms of getting more info for QA around dracut? Jul 22 12:58
jlaska | or should I leave it open to an interested tester Jul 22 12:59
Viking-Ice | I can ping harald about this sure Jul 22 12:59
Viking-Ice | try to gather some info Jul 22 12:59
Viking-Ice | etc.. Jul 22 12:59
jlaska | Viking-Ice: awesome, I appreciate it ... I spoke to harald briefly but I know you're well invested into the dracut/* wiki pages so it will make more sense to you! Jul 22 12:59
jlaska | #action Viking-Ice will check-in with Dracut maintainers to get additional feedback to guide testing Jul 22 13:00
jlaska | okay changing gears now ... Jul 22 13:00
jlaska | #topic AutoQA update Jul 22 13:00
--- | zodbot has changed the topic to: AutoQA update (Meeting topic: Fedora QA meeting) Jul 22 13:00
jlaska | wanted to give wwoods and dpravec a chance to talk about updates on the autoqa space Jul 22 13:00
jlaska | wwoods: you want to go first? Jul 22 13:00
jlaska | while we wait for wwoods ... Jul 22 13:02
jlaska | dpravec: I know you got to feel the rawhide burn last week Jul 22 13:02
dpravec | i was talking to some people in brno today, and some of them think that we should go work on beaker Jul 22 13:02
dpravec | that autotest might be just waste of time in the long run Jul 22 13:03
jlaska | yeah it's a bit confusing Jul 22 13:03
jlaska | but when beaker is live ... we'll just be using autotest with beaker Jul 22 13:03
jlaska | we've been in talks with the beaker developers to make sure that the solution we have now will be viable when beaker has a working implementation Jul 22 13:03
dpravec | my testing installation still have some issues, i am sorry i cant report much this time, should be much better from me on monday Jul 22 13:03
wwoods | sorry, back now Jul 22 13:07
wwoods | so I've automated the first 4 test cases on https://fedoraproject.org/wiki/QA:Rawhide_Acceptance_Test_Plan Jul 22 13:07
wwoods | and have them running under autotest Jul 22 13:07
jlaska | dpravec: okay ... let's walk through the autotest packaging issues tomorrow so we can get f13 some feedback in the ticket Jul 22 13:07
dpravec | jlaska: thanks, i will need few more hours to make it Jul 22 13:08
wwoods | https://fedorahosted.org/autoqa/browser/rawhide_acceptance Jul 22 13:08
wwoods | working on the installation tests now Jul 22 13:08
jlaska | dpravec: you can get your system back to an F11 laptop ... and we can go from there with the virt stuff Jul 22 13:08
adamw | wwoods: are the results of the first automated tests going anywhere public yet? Jul 22 13:09
dpravec | yes i already did, my laptop is becoming useable again Jul 22 13:09
wwoods | adamw: no. Jul 22 13:09
wwoods | they'll go public eventually Jul 22 13:09
adamw | k Jul 22 13:09
wwoods | if you want to run the test yourself it works outside autotest Jul 22 13:09
wwoods | right now I'm just focusing on actually automating the agreed-upon test cases Jul 22 13:10
wwoods | getting the testing and results public will happen later Jul 22 13:10
--- | onekopak1 is now known as onekopaka Jul 22 13:11
jlaska | wwoods: cool stuff, I've got your updated tests packaged into autoqa and running on another private autotest server Jul 22 13:11
wwoods | jlaska: excellent Jul 22 13:11
wwoods | we're gonna need to set up separate ppc/x86_64/i386 hosts for some tests Jul 22 13:11
wwoods | not sure how we're going to test ppc in the absence of virt on those systems Jul 22 13:12
* | jlaska makes a note on ppc Jul 22 13:12
jlaska | it's possible ... just not using virt-manager/libvirt Jul 22 13:12
jwb | shame Jul 22 13:12
wwoods | yeah, it's not hard to start the test, just hard to recover when the test fails Jul 22 13:12
wwoods | when a virt guest dies, well, you're watching on its serial console Jul 22 13:12
wwoods | and you can just delete it and mark it failed Jul 22 13:13
wwoods | when an actual machine dies you need to automatically reinstall. somehow. Jul 22 13:13
jlaska | wwoods: cobbler now has this support ... which is what I use for normal install testing Jul 22 13:13
jlaska | that's a bit more formal than using just libvirt ... but it's an option in the future if we need it Jul 22 13:14
wwoods | I'm not enthusiastic about requiring a cobbler server to run the rawhide tests, but if that's the only way I'm sure we can make it work Jul 22 13:14
f13 | wwoods: how hung up are we on install testing? Jul 22 13:15
f13 | if we threw out install testing and went back to just repo level testing, could we have something useful quicker? Jul 22 13:15
jlaska | f13: yeah that's a possibility wwoods discussed a while back ... if it comes to it Jul 22 13:16
f13 | I feel we're falling short of "do something useful ASAP" Jul 22 13:16
f13 | and taking too much on again Jul 22 13:16
adamw | i dunno, it sounds like useful stuff is being done already Jul 22 13:17
jlaska | f13: hmm, we are slightly behind the schedule set forth ... but I'm not too concerned yet Jul 22 13:17
adamw | the four tests wwoods has automated already are repo-level testing Jul 22 13:17
wwoods | yes. Jul 22 13:17
f13 | adamw: yes, and we don't have them happening in public, automated, reportable way Jul 22 13:17
jlaska | all of the tests are public Jul 22 13:18
jlaska | and we have stuff "in the works" for getting the server and client systems open Jul 22 13:18
f13 | jlaska: the original schedule was 3 weeks after last FUDCon Boston (: Jul 22 13:18
jlaska | f13: the original schedule is here https://fedorahosted.org/autoqa/milestone/israwhidebroken.com Jul 22 13:18
jlaska | where we set a date of Aug 18 Jul 22 13:18
f13 | jlaska: that's the current iteration of this automated QA effort. Jul 22 13:19
jlaska | f13: right on, that's the plan we are working Jul 22 13:19
f13 | however this whole thing started back at FUDCon Boston with the idea of doing something useful in 3 weeks, useful in 3 months, useful in 3 quarters Jul 22 13:19
wwoods | that's great Jul 22 13:19
adamw | nothing we can do now is magically going to make that happen. Jul 22 13:19
f13 | so I feel that I've failed miserably in delivering something useful to our maintainers based on those timelines. Jul 22 13:20
adamw | the point is we actually started working on a concrete plan to do this stuff relatively recently, and as far as i've seen, progress has been pretty rapid Jul 22 13:20
wwoods | we have a plan. we are following the plan. if you would like to help, feel free. Jul 22 13:20
jlaska | f13: if you set the criteria for success at 3 weeks, then yes we missed that target Jul 22 13:20
f13 | wwoods: I've been trying to help, but the targets keep moving Jul 22 13:20
adamw | i just listen to the updates from wwoods each week, but from those he seems to be making solid strides each week Jul 22 13:20
f13 | I thought we were very close to having something useful with autoqa, then we threw that out for autotest and I lost a few weeks to trying to package that up Jul 22 13:21
f13 | (still not done) Jul 22 13:21
f13 | now we may be throwing that out for beaker? Jul 22 13:21
adamw | just a few weeks ago we had nothing, now we have a full set of test cases, a full test plan, and four automated tests; i think that's indicative of substantial progress. the fact that it took a long time to get to the starting gate is something we can't do anything about now Jul 22 13:21
wwoods | no. Jul 22 13:21
jlaska | f13: you're keeping it honest ... the israwhidebroken.com plan doesn't match up with the FAD dates Jul 22 13:21
jlaska | f13: nope we are not changing for beaker in the short term Jul 22 13:21
wwoods | beaker is inventory / system provisioning. it currently lacks a test harness and scheduler Jul 22 13:21
f13 | adamw: we had automatible tests months ago Jul 22 13:21
wwoods | autotest is a scheduler and test harness. it currently lacks provisioning capabilities. Jul 22 13:21
wwoods | as for the autoqa stuff Jul 22 13:22
wwoods | both autoqa tests we had before Jul 22 13:22
wwoods | are being run under autotest already Jul 22 13:22
wwoods | and now we're writing new tests Jul 22 13:22
f13 | this isn't directed at anybody, I'm just frustrated that we still don't have a simple tool doing a simple test and sending out a simple email Jul 22 13:23
wwoods | and actually having those results reported somewhere we can look at 'em Jul 22 13:23
f13 | something I felt we could have had 3 weeks after FUDCon if that's what we actually targeted Jul 22 13:23
jlaska | f13: I hear ya ... you have valid concerns. I can tell you now that we are on schedule with our plan and things are progressing Jul 22 13:23
wwoods | but yeah, it's not sending out email anymore Jul 22 13:23
f13 | I'm also frustrated because my window for helping out is rapidly closing, because Alpha is coming, then snapshots, then beta, then release Jul 22 13:24
jlaska | okay ... we can keep tabs on this, but let's close it out for today Jul 22 13:24
f13 | in fact, I don't think I have any time left, since I need to get signing server finished and maybe enact some of the FAD proposals Jul 22 13:24
jlaska | f13: that's okay ... the only action item you have is the autotest packaging help Jul 22 13:24
jlaska | and that's looking good ... and dpravec will have some more information there if needed Jul 22 13:25
--- | stickster_food is now known as stickster Jul 22 13:25
jlaska | if time runs out ... we can reassign that task if needed Jul 22 13:25
jlaska | okay ... diving into open discussion in a moment ... Jul 22 13:25
jlaska | #topic Open discussion Jul 22 13:25
--- | zodbot has changed the topic to: Open discussion (Meeting topic: Fedora QA meeting) Jul 22 13:25
jlaska | sdziallas: you mentioned back at the top of the meeting about plans around a 'sugar on a stick' test day Jul 22 13:26
* | mchua listening Jul 22 13:27
* | jlaska seeks to lubricate the discussion Jul 22 13:27
sdziallas | jlaska: yup, that's true. I've been doing the Sugar on a Stick work for Sugar Labs lately, which is basically a rebranded Fedora base system with the Sugar packages on top. We're really interested in maintaining a good connection to the Fedora community (given that I started as a Fedora guy, too)... and would also like to test the Sugar implementation in Fedora Jul 22 13:27
dpravec | i was talking with people and it seems to me that best what we can do to improve quality of releases is to make some smoother rawhide repository, which will contain only packages without serious bug or dependency problems. similar to debian unstable/testing... this would enourage people to test latest (or week old) packages. Jul 22 13:27
f13 | #topic Sugar on a stick testing Jul 22 13:28
f13 | d'oh (: Jul 22 13:28
jlaska | :) Jul 22 13:28
jlaska | #topic open discussion [sugar on a stick test day] Jul 22 13:28
--- | zodbot has changed the topic to: open discussion [sugar on a stick test day] (Meeting topic: Fedora QA meeting) Jul 22 13:28
f13 | dpravec: we're going to get essentially that if my no frozen rawhide proposal goes though Jul 22 13:28
jlaska | adamw: you and I discussed this a while back as well ... I suggested to sdziallas that it might make sense to use the test day SOP to host a sugar test day event Jul 22 13:28
jlaska | https://fedoraproject.org/wiki/User:Adamwill/Draft_test_day_SOP Jul 22 13:29
adamw | yes Jul 22 13:29
adamw | which i still haven't moved out of draft yet Jul 22 13:29
adamw | le sigh Jul 22 13:29
jlaska | sdziallas: is there information you need to help get things started with self-hosting? Jul 22 13:30
sdziallas | I think that might be an interesting way of realizing a Sugar / Sugar on a Stick test day... Jul 22 13:30
jlaska | sdziallas: we also mentioned some of the work Adam Miller is doing with XFCE weekly spins Jul 22 13:30
dpravec | another thing we can do is to provide tools for developers/testers to test LATEST or custom packages on some test servers using gui (vnc, nx...) Jul 22 13:30
jlaska | dpravec: one sec ... Jul 22 13:30
Viking-Ice | ? Jul 22 13:30
dpravec | just today i was talking to openjdk developer who would be happy to have this... Jul 22 13:30
jlaska | dpravec: I've got your points down ... just want to finish the sugar topic first Jul 22 13:31
sdziallas | jlaska: heh, well... we're looking to get some weekly rawhide spins for a sugar spin, too, so this might help. Jul 22 13:31
sdziallas | jlaska: I think it would be good to pick a date soonish to have something we can plan with... (so that we don't have two test days at the same date or so) Jul 22 13:32
adamw | schedule's at https://fedoraproject.org/wiki/QA/Fedora_12_test_days Jul 22 13:32
adamw | you can take any of the open slots on the left hand side Jul 22 13:32
sdziallas | adamw: well, if this is going to follow the SOP proposal, will it still be one of the usual test day dates then? Jul 22 13:33
adamw | don't see why not Jul 22 13:33
adamw | there's no reason test days on the main schedule have to be organized by members of the cabal ;) Jul 22 13:33
jlaska | if the expertise at hosting the event are outside of fedora QA proper ... we might wnat to use a different track Jul 22 13:33
jlaska | that was my only concern Jul 22 13:34
adamw | i don't think we need a zillion tracks Jul 22 13:35
sdziallas | adamw, jlaska: imo, something around (if not even that date) the September 3 slot would be great... Jul 22 13:35
adamw | a track makes sense for fit and finish because they have a sequence of events that are part of a clearly defined project Jul 22 13:35
adamw | this is a single event, aiui Jul 22 13:35
jlaska | adamw: track is a poor use of terms on my part Jul 22 13:35
adamw | makes sense to just stick it in the main schedule, for me Jul 22 13:35
jlaska | to me, that means Fedora QA owns the event, is that wrong? Jul 22 13:35
jlaska | owns in terms of coordinating, announcing, etc... Jul 22 13:36
adamw | i don't see that it has to. Jul 22 13:36
jlaska | sdziallas: is that stuff you and mchua can offer help in ... e.g. test plans/cases Jul 22 13:36
adamw | but i don't think it's terribly important, let's just stick a date on the damn thing and we can argue the semantics later Jul 22 13:36
jlaska | heh ... semantics Jul 22 13:36
adamw | sept 3 looks fine to me Jul 22 13:36
jlaska | adamw: so you're saying sept 3? Jul 22 13:37
jlaska | ;) Jul 22 13:37
jlaska | a'ight .. . we've got a date Jul 22 13:37
jlaska | I'll bring the flowers Jul 22 13:37
sdziallas | jlaska: we had some plan originally in the SugarLabs wiki, but I can certainly help (and maybe mchua or others, too) Jul 22 13:37
sdziallas | :) Jul 22 13:37
mchua | jlaska: if it would help to simultaneously use this as a live test of what we're trying to do with smw, that can be arranged Jul 22 13:38
jlaska | mchua: yeah that'd be a nice tie-in Jul 22 13:38
mchua | jlaska: we've got a concrete test case case to set up for on tuesday then Jul 22 13:38
jlaska | mchua: you got it Jul 22 13:39
jlaska | okay, so we'll follow-up as we get closer to Sept 3 Jul 22 13:39
sdziallas | awesome! Jul 22 13:39
jlaska | okay, back to dpravec ... Jul 22 13:39
jlaska | #topic open discussion [rawhide testing] Jul 22 13:39
--- | zodbot has changed the topic to: open discussion [rawhide testing] (Meeting topic: Fedora QA meeting) Jul 22 13:39
jlaska | dpravec: you had some ideas ... were there any things you wanted to walk through in the meeting? Jul 22 13:40
* | jlaska has to wrap things up here Jul 22 13:40
jlaska | stomach digesting itself Jul 22 13:41
jwb | jlaska, sounds like a hot new diet fad Jul 22 13:41
jlaska | jwb: no kidding Jul 22 13:41
dpravec | hmm users keep telling me that they really need to have useable fresh packages. having rawhide broken makes it hard for them to even try it Jul 22 13:41
* | adamw hasn't had his fricking breakfast yet Jul 22 13:41
adamw | usable and fresh are so often mutually exclusive. this is the problem with development branches. Jul 22 13:41
dpravec | this week is rawhide full of problems, uninstallable... broken Jul 22 13:42
Viking-Ice | the ususal Jul 22 13:42
jlaska | dpravec: I think the proposal f13 pointed to is aiming to address some of your concerns Jul 22 13:42
dpravec | yes i will read details of those and comment there if appropriate Jul 22 13:42
jlaska | please do, I know you have experience with the deb stable/unstable model Jul 22 13:42
dpravec | even debian experimental is more stable than rawhide is Jul 22 13:43
f13 | dpravec: the biggest change with my proposals is that we wouldn't even be making install images for rawhide Jul 22 13:43
jeff_hann | dpravec, I beg to differ Jul 22 13:43
f13 | we'd only make install images after we've branched after feature freeze Jul 22 13:43
f13 | so "rawhide" would just be a repo of packages, just like Debian unstable Jul 22 13:43
jeff_hann | I used Debian, all branches Jul 22 13:43
* | jlaska senses rat hole Jul 22 13:43
dpravec | jeff_hann:  :) Jul 22 13:43
dpravec | well i just tried rawhide on my desktop Jul 22 13:44
f13 | you'd have to start with something more stable, like the branched Fedora or the previous release, and then yum update to rawhide versions of packages Jul 22 13:44
dpravec | and compared to debian unstable, its unusable Jul 22 13:44
Viking-Ice | if they are going to run developing components on a stable product the package maintainer needs to setup up repo for example on his people page and build those developing package for F11 like the virtualization team did.. Jul 22 13:44
f13 | so we'd lose the expectation that rawhide is "installable" Jul 22 13:44
jeff_hann | compared to unstable, yes Jul 22 13:44
f13 | dpravec: that's an unfair comparison, because you can't directly install Debian unstable. You have to start out on something else, and then /update/ to unstable Jul 22 13:44
dpravec | f13: yes you can Jul 22 13:44
f13 | dpravec: I don't recall there being nightly install images made from Debian unstable Jul 22 13:44
dpravec | there are daily /weekly install images Jul 22 13:45
dpravec | and live images Jul 22 13:45
jlaska | okay gang ... I'm sensing a great discussion topic for after the meeting :) Jul 22 13:45
jlaska | let's call end of meeting Jul 22 13:45
f13 | and really, comparing us to debian is jokes Jul 22 13:45
Viking-Ice | or on next meeting.. Jul 22 13:45
f13 | different goals, different methodologies Jul 22 13:45
dpravec | anaconda is big problem, its very complex Jul 22 13:45
jlaska | and we can take follow-up to the debian/rawhide discussion to hte list or #fedora-qa Jul 22 13:45
dpravec | f13: ic :) Jul 22 13:45
jeff_hann | :) Jul 22 13:45
f13 | right, anaconda is complex, because it has lots of useful features Jul 22 13:45
dpravec | i would love to get to the situation that users will not be afraid to install fedora 'unstable' Jul 22 13:46
f13 | testing those features before they're ready to be tested is a mistake and useless Jul 22 13:46
f13 | so lets not do it Jul 22 13:46
jlaska | thanks folks for your time ... today was quite a long one! Jul 22 13:46
f13 | dpravec: I'd love to get to the situation where the only install images we provide of Fedora 'unstable' are ones that we know are ready and useful to be tested Jul 22 13:46
jlaska | #endmeeting Jul 22 13:46
f13 | otherwise you start where we know things work, and update to 'unstable' from there Jul 22 13:46
--- | zodbot has changed the topic to: Channel is used by various Fedora groups and committees for their regular meetings | Note that meetings often get logged | For questions about using Fedora please ask in #fedora | See http://fedoraproject.org/wiki/Meeting_channel for meeting schedule Jul 22 13:46
zodbot | Meeting ended Wed Jul 22 17:46:52 2009 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . Jul 22 13:46
zodbot | Minutes: http://meetbot.fedoraproject.org/fedora-meeting/2009-07-22/fedora-meeting.2009-07-22-16.02.html Jul 22 13:46
zodbot | Minutes (text): http://meetbot.fedoraproject.org/fedora-meeting/2009-07-22/fedora-meeting.2009-07-22-16.02.txt Jul 22 13:46
dpravec | yes Jul 22 13:46
zodbot | Log: http://meetbot.fedoraproject.org/fedora-meeting/2009-07-22/fedora-meeting.2009-07-22-16.02.log.html Jul 22 13:46
dpravec | we need ready, usefull and tested images of unstable daily ! Jul 22 13:47
dpravec | lol Jul 22 13:47
jlaska | dpravec: f13 jeff_hann: feel free to continue ... but I've got to step out Jul 22 13:47
jeff_hann | see ya :) Jul 22 13:47
jlaska | I'll follow-up to the list with minutes later today Jul 22 13:47
jlaska | thanks folks! Jul 22 13:48
dpravec | if not with latest anaconda, thats ok... Jul 22 13:48
jeff_hann | guys, I don't think that comparisons to Debian's unstable are really fruitful and useful Jul 22 13:48
f13 | dpravec: no we don't. Jul 22 13:49
f13 | dpravec: install images are only useful for testing the install process Jul 22 13:49
dpravec | f13 why? Jul 22 13:49
f13 | dpravec: if the install process isn't ready to be tested, then why provide images to test it? Jul 22 13:49
dpravec | the idea is to make it ready... and make it tested Jul 22 13:49
f13 | and until it's ready, throwing it at people is worse than worthless Jul 22 13:49
dpravec | i believe we can do it, at least partially Jul 22 13:50
jeff_hann | you can't have it ready without it being extensively tested Jul 22 13:50
f13 | wrong Jul 22 13:50
dpravec | my first questions when i came to qa team was: where can i take todays live image of rawhide? Jul 22 13:50
jeff_hann | why? Jul 22 13:50
dpravec | my teammate is asking the same Jul 22 13:50
f13 | ready to be tested is vastly different than ready to be a GOLD release Jul 22 13:50
jeff_hann | obviously Jul 22 13:50
f13 | dpravec: why did you want that? Jul 22 13:50
f13 | dpravec: what was your end goal for having that? Jul 22 13:50
dpravec | to run it and to watch how it works Jul 22 13:51
f13 | dpravec: install F11, update to rawhide Jul 22 13:51
dpravec | yes that waht everyone keeps telling me Jul 22 13:51
f13 | because that's what works, and would be of useful feedback Jul 22 13:51
f13 | and that's whats ready to be tested Jul 22 13:51
dpravec | i even tried last friday, ended with unusable computer - could not get even to single mode Jul 22 13:52
f13 | so that's not an installer issue, thats an issue with some other package in the set Jul 22 13:52
dpravec | (and i reinstalled it with encrypted home this time) Jul 22 13:52
f13 | unless you mean you couldn't even install F11 Jul 22 13:52
dpravec | i would be happy to have something like debian 'testing' Jul 22 13:53
dpravec | having only packages from rawhide which have no critical bugs Jul 22 13:54
dpravec | for 14 days Jul 22 13:54
f13 | and how do you propose to discover those packages? Jul 22 13:54
dpravec | 1 ) automated tests Jul 22 13:54
f13 | and even if you did, how would you get anybody to test things in unstable? Jul 22 13:54
dpravec | 2) users reporting deadly problems Jul 22 13:54
dpravec | f13: and it is possible to test rawhide now? Jul 22 13:55
dpravec | i have seen 3 people this week trying to install it somehow Jul 22 13:55
f13 | great, we can revisit this once we have a comprehensive set for #1 Jul 22 13:55
dpravec | and all failed Jul 22 13:55
dpravec | even upgrade from f11 Jul 22 13:55
f13 | dpravec: for the vast majority of packages in rawhide yes. Jul 22 13:55
f13 | dpravec: start with F11, update to the specific package you want to test Jul 22 13:55
dpravec | f13 i agree i am talking about future, but not that far away Jul 22 13:55
dpravec | this year maybe Jul 22 13:55
f13 | our package interdependence doesn't lend itself very well to trying to arbitrarily break up package versions into layered repos Jul 22 13:56
f13 | you'd often have one package, that has reached it's magical 14 day state, you try to put it in testing, but it has to drag along stuff that hasn't reached that 14 day mark Jul 22 13:56
dpravec | another way to move package (even faster) is to have users testing it Jul 22 13:57
dpravec | i mean not only install and file bugs Jul 22 13:58
dpravec | but really test, maybe on test servers via vnc Jul 22 13:58
dpravec | at least some basic tests Jul 22 13:58
dpravec | its installable, it runs, it seems ok Jul 22 13:58
dpravec | making this automated would be good Jul 22 13:58
dpravec | but some parts are hard to automate Jul 22 13:59
f13 | yeah, I'm totally willing to revisit these ideas when we have that Jul 22 13:59
dpravec | ok Jul 22 13:59
dpravec | will try to focus on autotest for now Jul 22 14:00
dpravec | :) Jul 22 14:00
dpravec | when we will see some autotest results and numbers, it could help us to make good decisions Jul 22 14:00
jeff_hann | good night, guys Jul 22 14:01
jeff_hann | take care :) Jul 22 14:01
dpravec | good night Jul 22 14:01

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!