fedora-qa
MINUTES
16:00:50 <adamw> #startmeeting Fedora QA meeting
16:00:50 <zodbot> Meeting started Mon Dec 15 16:00:50 2014 UTC.  The chair is adamw. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:50 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:00:54 <adamw> #meetingname fedora-qa
16:00:55 <zodbot> The meeting name has been set to 'fedora-qa'
16:00:59 <adamw> #topic Roll call
16:01:03 * roshi is here
16:01:09 <adamw> ahoyhoy folks, who's here for another week of exciting qa?
16:01:15 * pschindl is here
16:01:45 * satellit_e listening
16:04:12 <adamw> man, thin turnout
16:04:12 <brunowolff> I'm going to try to follow the discussion.
16:04:15 <adamw> .fire everyone
16:04:15 <zodbot> adamw fires everyone
16:04:41 <pschindl> kparal is on PTO
16:05:29 <adamw> pfeh, vacations.
16:05:35 <roshi> tflink is around, probably getting coffee or something
16:05:41 * tflink is here
16:05:47 <roshi> I mean, he just ran the qa-devel meeting
16:05:49 <adamw> first agenda item: all vacations and sleep are banned from here on out.
16:06:08 <roshi> is that an AND or XAND?
16:06:30 * nirik is lurking, but reading the email firehose.
16:06:32 <adamw> it's an AND, AND a .fire roshi for insubordination
16:06:42 <roshi> haha
16:06:46 <adamw> (but no, this does not mean you can sleep)
16:06:48 <brunowolff> Hey, I get more time to work on Fedora when I'm on vacation.
16:06:53 * mkrizek joins
16:07:14 <adamw> brunowolff: in that case...PERMANENT VACATION
16:07:19 <tflink> adamw: but how would you know if we're sleeping if it looks like we're working?
16:07:34 * tflink looks into a markov-chain irc bot
16:07:42 <adamw> well, yes, that's true. the quality of your output in both states is difficult to tell apart. ;)
16:08:07 <tflink> I didn't realize that the quality of my output when I was sleeping was so high
16:08:08 <adamw> allllrighty then, that's enough light joshing, according to the Big Meeting Manual
16:08:13 <roshi> he's got santa wired into his network - he knows when we're sleeping and when we're awake, and other things about us
16:08:14 <adamw> haha, nicely saved.
16:08:15 * smccann_ is here
16:08:27 <adamw> roshi: think about it: have you ever seen me and santa in the same room?
16:08:34 <adamw> i'll just leave you with that.
16:08:38 <adamw> #topic Previous meeting follow-up
16:08:42 <roshi> lol
16:08:49 <tflink> adamw == santa?
16:09:12 <adamw> my lips are HO HO HO sealed.
16:09:16 * roshi 's head asplode
16:09:34 <adamw> "roshi to look at implementing a compose event listener in taskotron" - roshi, did you get any time for that?
16:09:51 <roshi> I didn't, but tim indicates that martin already wrote it
16:09:53 <tflink> review just posted for that code, waiting for more testing to make sure it works
16:09:55 <roshi> aiui
16:10:00 * adamw adds roshi to naughty list
16:10:07 <tflink> https://phab.qadevel.cloud.fedoraproject.org/D268
16:10:11 * adamw adds tflink to slightly-less-naughty list
16:10:15 <adamw> tflink: awesome
16:11:24 <adamw> #info mkrizek has written a compose listener already, it is pending review at https://phab.qadevel.cloud.fedoraproject.org/D268
16:11:40 <adamw> oops
16:11:41 <adamw> #undo
16:11:41 <zodbot> Removing item from minutes: INFO by adamw at 16:11:24 : mkrizek has written a compose listener already, it is pending review at https://phab.qadevel.cloud.fedoraproject.org/D268
16:12:03 <adamw> #info "roshi to look at implementing a compose event listener in taskotron" - mkrizek has written a compose listener already, it is pending review at https://phab.qadevel.cloud.fedoraproject.org/D268
16:12:58 <adamw> #info "adamw to work on wiki magic and relval [for nightly compose testing]" - done and released/deployed, https://lists.fedoraproject.org/pipermail/test-announce/2014-December/000969.html
16:13:12 <adamw> anyone have any follow-up or thoughts on that stuff, or shall we move on?
16:13:33 * tflink is planning to work on the package reviews for that this week
16:13:39 <tflink> unless someone else gets to them first
16:13:42 <adamw> thanks
16:14:33 <adamw> #topic Tooling check-in: taskotron, blockerbugs, relval, etc
16:14:38 <adamw> this is one that got left over from last week
16:15:02 <adamw> just thought it'd be good to sync up and update status of tooling work for the f22 cycle, for folks not following the qa-devel meetings
16:15:17 <tflink> we ended up postponing the qa-devel meeting until tomorrow due to some last-minute absences
16:15:50 <tflink> several proposals have been posted to qa-devel@ in the last week or so. there has been a little discussion but not much so far
16:16:20 <tflink> as usual, we have way too much potential work to get done before f22 branch (mid-february?)
16:16:26 <tflink> so we'll be setting some priorities
16:16:52 <tflink> fedmsg integration is going to be a relatively high priority since bodhi2 is expected to go into production sometime in january/february
16:17:06 * roshi looks forward to seeing that :)
16:17:45 <tflink> beaker is a bit of a wild card. still not sure if/when that will go to a production system but we're waiting on a RFE that is expected to land in January some time
16:18:29 <adamw> what do you mean by fedmesg integration exactly? sending fedmsg messages?
16:18:47 <tflink> yeah, sending fedmsgs on task completion instead of creating bodhi comments
16:19:21 <tflink> bodhi2 is capable of displaying data from resultsdb for a given update and we don't need to be making comments for feedback anymore
16:19:43 <tflink> but without those comments, we lose the ability to notify maintainers of status
16:20:05 <tflink> we're planning to move those notifications over to fedmsg + fmn
16:21:11 <tflink> another open question is whether we want to invest much/any time improving blockerbugs
16:21:21 <tflink> if we do that, it'll come at the cost of Taskotron features
16:21:50 <roshi> adamw is a developer now, couldn't he work on blockerbugs?
16:21:52 * roshi ducks
16:21:54 <roshi> :p
16:22:35 * tflink certainly isn't going to stop anyone from taking on blockerbugs work
16:22:38 <adamw> you're lucky i fired you already
16:23:04 <roshi> haha
16:23:15 <tflink> adamw: because that's ever stopped you from firing people multiple times in one day?
16:23:17 * tflink ducks
16:23:19 <adamw> i guess my thought would be if we can get disposable clients done in time to actually do some f22 testing with it, that'd be high priority
16:23:26 <adamw> tflink: i'm trying to be nice. it's christmas!
16:23:39 <tflink> fair enough :)
16:23:40 <adamw> if not, maybe blockerbugs gets a bit more interesting. but it's always a hard call
16:24:08 <tflink> it's going to be close if we decide to go for disposable clients for f22
16:24:14 * adamw has 'look at blockerbugs and see if you can do anything useful' on his extremely long list of discretionary tasks but doesn't mean i'll ever get there :/
16:24:22 <adamw> #chair tflink roshi
16:24:22 <zodbot> Current chairs: adamw roshi tflink
16:24:24 <tflink> there's not much time left, with folks disappearing, and all
16:24:33 <roshi> yeah
16:24:43 <tflink> ~ 1 month to get it all working
16:25:25 <tflink> which isn't impossible but isn't really a sure thing
16:26:49 <tflink> beaker has some interesting potential if we can get it working in time
16:26:54 <adamw> of course we don't even really have the f22 schedule yet, do we? it's still 'no earlier than'
16:26:56 <tflink> speficially, basic install tests
16:27:08 * roshi has to look up what all features blockerbugs wants...
16:27:09 * tflink is assuming worst case scenario right now
16:27:26 <adamw> hell, if it can do as much as 'run a kickstart and check if the install succeeded' we can do a lot of interesting things with that.
16:27:48 <tflink> adamw: it was doing that last week, something broke in the meantime and need to investigate
16:28:12 * satellit_e this is f22 workstation 20141215  bare metal
16:30:18 <adamw> tflink: well, that sounds relatively promising, though there's a lot of devils in the details of course
16:30:30 <adamw> satellit: cool
16:30:37 <tflink> adamw: aren't there always?
16:30:48 <satellit_e> for testing page?
16:30:50 <roshi> I thought that's where all devils lived
16:32:30 <adamw> satellit: 1215 doesn't have any different installer-related packages from 1208, so i didn't set up pages for it...i'll see if bcl is planning new builds this week
16:32:35 <adamw> roshi: hehe
16:33:10 <adamw> ok, so aside from fedmsg, disposable clients, and blockerbugs, what are the other development 'opportunities'?
16:33:44 <roshi> I'm working more on testCloud for cloud testing
16:34:16 <roshi> but that's just me, not a qa-devel thing :)
16:34:39 <jreznik> adamw: right, we don't have f22 schedule, only changes are scheduled so far...
16:34:59 <tflink> jreznik: any idea when that might be decided?
16:35:18 <tflink> adamw: those are the big ones right now
16:35:22 <adamw> ok
16:35:36 <tflink> we want to get qadevel into fedora infra, but I don't see that happening right now
16:35:50 <tflink> it'll require a bunch of ansible work and packaging phabricator for fedora/epel
16:35:52 <adamw> #info major tooling development directions for F22 cycle are fedmsg integration for taskotron (sending out fedmsg messages), disposable test clients for taskotron, and blockerbugs
16:36:01 <jreznik> tflink: it's usually decided once we have all changes in hands but this time, there's also some serious push to change how release is scheduled :( so no answer, sorry
16:36:02 <adamw> OK
16:36:33 <tflink> jreznik: no worries, we've kinda gotten used to that unknown
16:36:50 <tflink> still makes planning difficult, but it's not a new thing
16:36:51 <adamw> i don't really have anything exciting for relval unless anyone had thoughts/suggestions/whatever, just thew it on the end of the list. mostly i'm aiming to do a lot of cleanup on it, and write some tests.
16:37:01 <adamw> also, god, unicode.
16:37:16 <tflink> adamw: you could port to python3
16:37:29 <adamw> tflink: first i'd have to port python-mwclient to it
16:37:38 <adamw> which is also on the long list of discretionary things...
16:37:52 <tflink> sounds like my list :)
16:37:56 <adamw> i think so :)
16:38:00 * tflink would love to see a python3 koji client
16:38:46 * pschindl has to leave
16:39:01 <adamw> so i guess perhaps the initial goal is to get disposable clients done, but if it's not looking possible, we can maybe diversify a bit and do some blockerbugs work if there seem to be sufficiently useful changes possible
16:39:14 <adamw> cya pschindl! did you have any thoughts on the later topics that you want to leave?
16:39:25 <tflink> sounds good to me
16:40:17 <tflink> it's usually mkrizek or myself working on blockerbugs and we'd be pretty much critical to getting disposable clients working
16:40:24 <adamw> rgr
16:40:56 <tflink> getting an el7 build working is on the list of less-optional things
16:41:09 <tflink> but that doesn't involve many (if any) code changes
16:43:35 <adamw> ok, moving on while we have time...
16:43:45 <adamw> #topic Release criteria changes (esp. multiboot)
16:44:03 <adamw> again we're without cmurf unfortunately, but basically we need to at least look at the criteria issues that came up around the end of f22
16:44:13 <adamw> looks like the multiboot criteria need re-writing
16:44:32 <adamw> anyone feel like volunteering for that, or it going to be muggins again? :)
16:45:04 * roshi inches away slowly
16:45:37 <roshi> multi-boot criteria aren't an easy thing to nail down, and harder to make the code work in all the possible permutations
16:45:58 <roshi> though, if we do the anaconda alpha freeze thing, it might become easier
16:47:21 <roshi> since there would be more time to iron out the bugs with different multi-boots
16:49:11 <adamw> yeah, it's mostly a case of 'adjust for real world possibilities' in this case i think
16:49:18 <adamw> have to go back and refresh my memory, though
16:49:26 <roshi> same here
16:53:09 <adamw> #action adamw to sync with cmurf and look at revising multiboot criteria
16:53:18 <adamw> #topic Upgrade path discussion
16:53:41 <dgilmore> tflink: we have on the radar to port koji to python3
16:53:48 <dgilmore> tflink: it is a huge undertaking
16:53:54 <adamw> so i figured i'd throw this on here since it was quite a passionate discussion, though i'm not sure what the best way forward is and i don't know if mschwendt or ralf is around
16:54:20 <tflink> dgilmore: yeah, that's why my emphasis was on the koji client
16:54:37 * tflink thought that the client could be ported with much less effort than koji itself
16:54:43 <dgilmore> tflink: to get the client ported the library that is the bulk of koji needs ported
16:55:09 <tflink> dgilmore: ok, I hadn't looked into it much. thought that was mostly an xmlrpc client
16:55:41 <dgilmore> tflink: the python library is shared amongst everything and is the bulk of things
16:56:50 <adamw> sounds like none of the people who *really care a lot* about upgrade path issues are around. :P
16:57:56 <adamw> #info no-one with a side in the upgrade path issue is around for the meeting
16:57:58 <adamw> #topic Open floor
16:58:02 <adamw> as you were :)
17:00:32 <adamw> alrighty, sounds like we're about done
17:00:45 * roshi doesn't have anything
17:00:53 * tflink neither
17:00:59 <roshi> I'll be AFK wednesday/thursday
17:01:04 <roshi> probably
17:01:19 <roshi> depends on weather, really
17:02:03 <adamw> #info the first Heroes of Fedora Testing post for F21 is up at http://fedoramagazine.org/heroes-of-fedora-qa-fedora-21-2/ , thanks roshi
17:02:08 <adamw> any news on the second?
17:02:30 <roshi> I figure this evening or tomorrow
17:02:41 <roshi> and then the last on Friday or the following Monday
17:02:52 <adamw> coolbeans
17:02:59 <adamw> and then the call for test days ;)
17:04:26 <roshi> the Heroes post has 580-some views - which is nice
17:05:29 * adamw blogs a link to it
17:06:53 <adamw> alrighty, time to blow this sucker up
17:06:57 <adamw> thanks for coming, folks
17:07:33 <roshi> np - thanks for running it!
17:07:55 <tflink> adamw: thanks for running the meeting
17:08:09 <adamw> #endmeeting