16:00:07 <jlaska> #startmeeting Fedora Quality Assurance Meeting
16:00:23 <jlaska> #topic gathering essential carbon units
16:00:33 <onekopaka> carbon units?
16:00:43 <jlaska> aka people :)
16:00:44 <onekopaka> we're carbon units now?
16:00:53 <jlaska> hehe, we've always been
16:01:11 <onekopaka> but we're being refered to as carbon units?
16:01:19 <onekopaka> that's just not nice.
16:01:39 <jlaska> We'll be short a few people today, including adamw, dpravec, kparal, lili and rhe
16:01:54 <jlaska> #topic gathering essential 'people'
16:01:59 <jlaska> onekopaka: there ya go :)
16:01:59 <onekopaka> I'll be here for the fun of it.
16:02:26 <jlaska> onekopaka: okay
16:02:35 <jlaska> show of hands for the QA meeting
16:02:42 <jlaska> wwoods: Viking-Ice: you guys around?
16:02:48 * poelcat here
16:02:50 * onekopaka is looking for the rear USB ports on his ex-server
16:02:58 <jlaska> poelcat: welcome
16:04:04 * wwoods around
16:04:11 <jlaska> wwoods: hey there
16:04:26 <jlaska> Viking-Ice_: there he is :)
16:04:48 <jlaska> okay let's get started with a fairly quick review of last week
16:04:59 * jlaska operating out of proposed agenda at https://fedoraproject.org/wiki/QA/Meetings/20090803
16:05:10 <jlaska> #topic previous meeting review
16:05:20 * f13 
16:05:24 <jlaska> f13: welcome!
16:05:35 <jlaska> most of this stuff has been done so this should be quick
16:05:48 <jlaska> # [jlaska] - file bug against distribution for rawhide pungi problems
16:05:57 <Viking-Ice_> yup.. dont know how long I'll stay thou need to remove 3d tagets from the archery track ( group supposed to meet sometimed between 16:00 & 17:00 )
16:05:59 <jlaska> done and thanks to f13 to for helping resolve the issue
16:06:06 <jlaska> for the logs, see bug#514062
16:06:08 <buggbot> Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=514062 medium, low, ---, jkeating, CLOSED RAWHIDE, Rawhide missing images
16:06:19 <jlaska> Viking-Ice_: okay, thanks for the heads up
16:06:32 <jlaska> a few action items around the blocker bug event last week ...
16:06:33 <jlaska> # [poelcat] - to send blocker bug meeting announce
16:06:46 <jlaska> thanks poelcat ... I've got the link in the logs as https://www.redhat.com/archives/fedora-test-list/2009-July/msg00655.html
16:07:44 <jlaska> next up ...
16:07:45 <jlaska> # [adamw] - to summarize blocker meeting
16:08:03 <jlaska> I didn't see a summary to the list on this yet, but the meetbot logs are available ... http://meetbot.fedoraproject.org/fedora-bugzappers/2009-07-31/fedora-bugzappers.2009-07-31-15.01.html
16:08:41 <jlaska> also from the meeting
16:08:44 * poelcat can't figure out why we didn't put a blocker bug meeting on the schedule for 8/7
16:08:44 <jlaska> dpravec kept us entertained ...
16:08:48 <jlaska> # [dpravec] - posting some fun ascii art for blocker bug meeting
16:08:57 <jlaska> poelcat: yeah, I've got that in the agenda later on ... was going to ask you about it
16:09:18 <jlaska> Next up ... we wanted to post wwoods solution to rawhidewatch for the rpm XZ payload issue
16:09:21 <jlaska> # [jlaska] - talk to warren about a rawhidewatch post on rpm xz workaround for rawhide testers
16:09:28 <jlaska> thanks warren!  http://rawhidewatch.wordpress.com/2009/07/31/unable-to-update-to-rawhide-rpmlibpayloadisxz
16:09:38 <jlaska> And last item from last week ...
16:09:40 <jlaska> # [kparal] - investigate updates to Test Day livecd-creator instructions
16:09:53 <warren> jlaska: you didn't get back to me regarding wordpress account
16:09:59 <jlaska> kparal isn't here today either, but posted updates to the test list for review https://www.redhat.com/archives/fedora-test-list/2009-July/msg00549.html
16:10:00 <warren> jlaska: I can grant other wordpress accounts post access to rawhidewatch
16:10:04 <jlaska> warren: I did ... it's just 'jlaska'
16:10:06 <warren> oh
16:10:07 <warren> ok
16:10:23 <jlaska> warren: thanks for reminder though, I need to revisit wordpress ;)
16:10:41 <warren> jlaska: what e-mail was used withi that account?
16:11:06 <jlaska> warren: Not sure, I'll have to follow-up with you on that
16:11:12 <warren> ok
16:11:49 <jlaska> any other updates from last week that I missed?
16:11:57 * jlaska preparing to dive into agenda
16:12:48 <jlaska> #topic Dracut rawhide test update
16:13:04 <jlaska> Viking-Ice_: I've got a note here to keep this on the agenda in case you had any updates to share
16:15:21 * jlaska suspects we lost Viking-Ice
16:15:46 <Viking-Ice_> yup still here with updates
16:16:04 * Viking-Ice_ middle of fixing arrows :/
16:16:51 <jlaska> oh cool, any news to share?  Sounds like you're in the middle of a competition too ... want to hold off for mail list follow-up?
16:18:11 <Viking-Ice_> No let's just finished this ( had to finish this arrow :) )
16:18:34 * Viking-Ice_ sneaks my topic to the top
16:18:48 <jlaska> okay, you've got the floor ... go for it
16:18:49 <Viking-Ice_> So I emailed harald
16:21:03 <Viking-Ice_> And he informed me that him Dave Jones Peter Jones ( kernel ) and Hans ( Anaconda ) are making the switch in rawhide
16:22:15 <jlaska> Viking-Ice_: so that should be landing soon?
16:22:16 <Viking-Ice_> and the target are for test days is more root on a network drive and complex raid/lvm setups.
16:22:24 <jlaska> right on
16:22:38 <Viking-Ice_> so the most common setups should just work
16:22:44 <f13> they were looking into it, then ran into a big hurdle
16:22:56 <f13> regarding the buildrequires of the kernel, and I'm not sure if it actually got switched on or not
16:24:21 <jlaska> Viking-Ice_: any action needed in to short-term when this change occurs?
16:24:24 <Viking-Ice_> Did not come with the email the progress on the switch still feel like we are missing some are ( kernel/anaconda + something else ) well maybe it's just me
16:24:39 <jlaska> f13: the hurdle was for the network storage cases?
16:24:59 <Viking-Ice_> jlaska: nothing that got mentioned but I will stay in touch with Harald with regards to the test days and test case writing
16:25:00 <f13> jlaska: basically all firmware that's packaged outside the kernel, that would be interesting in the initrd
16:25:14 <f13> network storage, network cards, scsi cards
16:25:17 <jlaska> oh hmmm
16:25:50 <jlaska> Viking-Ice_: okay thank you
16:26:08 <jlaska> so has this change already happened?
16:26:16 <jlaska> since we are past feature freeze now
16:27:06 * jlaska sees mkinitrd on rawhide installed system
16:27:31 <Viking-Ice_> dont think so since I wrote and he responded to the letter on friday and he says they are making the switch together
16:27:39 <jlaska> Viking-Ice_: okay
16:27:59 <jlaska> Viking-Ice_: I'll follow-up with you later on so I can come up to speed also
16:28:08 <Viking-Ice_> Ok
16:28:16 <jlaska> Alrighty ... let's change over to the Alpha ...
16:28:21 <jlaska> #topic Alpha test compose
16:28:34 <f13> yeah, this one is still MIA
16:28:44 <f13> due to issues blocking installability
16:28:48 <jlaska> Looks like the udev integration was resolved late last week (thanks clumens+jeremy)
16:28:54 <f13> sortof
16:28:59 <f13> the patch got naked on list
16:29:01 <jlaska> oh?
16:29:06 <f13> at least the first one did
16:29:09 <jlaska> on anaconda-devel-list?
16:29:15 <f13> as it broke all the blacklisting of devices and such
16:29:35 <jlaska> so what's in rawhide now then?
16:29:43 <jlaska> it's working around the reported problem
16:30:26 <f13> not sure.  Clumens posted a patch, jeremy meh'd it in, and then bill naked it, clumens asked bill for a better patch, bill posted some suggestions, and I haven't seen any traffic regarding it since then.
16:30:38 <denise> I think Jeremy and clumens got one inon Friday
16:30:45 * jlaska thought so too
16:31:04 <denise> and basically reviewed ach other
16:31:14 <jlaska> Word I have from clumens is that the anaconda-12.7 build had a fix to work around the issue
16:31:21 <jlaska> http://git.fedorahosted.org/git/?p=anaconda.git;a=commit;h=e6402846cae226519bfc34feb4f92b40d3bf0885
16:31:43 <f13> Yeah I'm looking at commit list now and I see jeremy committed something pretty awful
16:31:46 <denise> see jeremy's mail to anaconda-devel "Fix up udev sillies"
16:32:03 <jlaska> f13: I think that unwedges the test blockers that I was aware of
16:32:11 <denise> f13: yeah, noone's happy about this but had to get rawhide limping
16:32:40 <f13> I missed jeremy's patch on anaconda-devel-list (doesn't seem to be there)
16:32:45 <f13> so I thought it was still in limbo
16:32:57 <Viking-Ice_> got to run later...
16:32:57 <denise> f13: justthe commit mail - he had clumens review over his shoulder
16:32:59 <f13> jlaska: if that's the case, I'll bank out a test compose as soon as all my freaking meetings are over.
16:33:05 <f13> denise: gotcha.
16:33:13 <f13> Bill is OK with this too?
16:33:16 <jlaska> f13: okay ... I'll update that releng ticket with bug status
16:33:17 <f13> avoids the concerns he raised on list?
16:33:19 <denise> AFAIK
16:33:26 <denise> does avoid the blacklist problem
16:33:35 <jlaska> f13: when the compose is available, can you drop a note in that ticket and lili will commence testing
16:33:43 <f13> yep
16:33:47 <jlaska> thanks!
16:34:42 <skvidal> f13: jeremy did a test compose on friday with pycurlgrabber included - and everything worked :)
16:34:49 <f13> *cough*
16:35:53 <jlaska> Okay, so that's all I had on the Alpha test compose ... it'll be 4 days late ... but we'll get started walking the matrix now that the test blockers have been removed
16:36:26 <jlaska> Well, the other question I wanted to raise on this given the alpha test compose slip ... how does this impact the alpha
16:36:47 <jlaska> I think the best way to answer that now is that we are behind on testing media by 4 days
16:37:10 <jlaska> I don't think it says slip yet ... but it's a good indicator if we begin to miss other targets
16:37:37 <f13> sure, now that the blockers are cleared, a walking of the matrix outside what you expect to do with media would be /really/ good
16:37:43 <f13> to figure out what else is waiting for us
16:38:00 <jlaska> right on
16:38:19 <jlaska> Liam will kick that off tonight and then the fun begins!
16:38:43 <f13> what's your definition of 'tonight' ?
16:38:48 <f13> isn't rawhide available now for such testing?
16:39:13 <jlaska> indeed ... when Liam wakes up ... he'll start the ball rolling with the test matrix
16:39:20 <f13> ah
16:39:26 <f13> forgot what tz liam is in
16:39:26 <jlaska> but of course, people can download and test rawhide now ... and post results later
16:39:52 <jlaska> okay, if not other questions ... changing over to another Alpha topic
16:40:14 <jlaska> s/not/no/
16:40:22 <jlaska> #topic Alpha blocker bug day#4
16:40:34 <jlaska> looking at the schedule, I don't see an alpha blocker bug day scheduled for this Friday
16:40:48 <jlaska> given the test compose delay, I think it might be in order and wanted to raise the topic
16:41:02 * poelcat will add one to the schedule
16:41:03 <jlaska> poelcat: f13: what's your take?
16:41:21 <f13> My take was that releng was going to be doing one anyway to make sure we can compose the release candidate
16:41:33 <f13> so doing one on friday, just to do another one on Monday I think seemed silly
16:41:47 <f13> at least thats what I remember of the reasoning during our big schedule meeting
16:42:08 <poelcat> seems like a good idea if we are making "go/no go" following monday
16:42:13 <poelcat> #20 http://poelstra.fedorapeople.org/schedules/f-12/f-12-releng-tasks.html
16:42:24 <f13> we would have to look at them /again/ on monday
16:42:44 <jlaska> yeah, what's the worst that happens
16:42:45 <f13> I'm not against doing it on Friday, I just want everybody to understand that we'd be doing it on two successive work days
16:42:47 <poelcat> not really... if they aren't fixed or in modified then not ready?
16:42:50 <jlaska> there is no movement and the monday meeting is quick
16:43:45 * jlaska likes consistency ...
16:43:45 <poelcat> f13: so "go/no go" was really a "blocker review meeting" vs. a quick meeting to decide if ready to go or not
16:43:59 * poelcat wonders about changing task name
16:44:07 <poelcat> ?
16:44:11 <f13> poelcat: well, in order to give a go/no go, we have to look at the blocker list
16:44:16 <f13> essentially doing a blocker review
16:44:18 <jlaska> let's split it out
16:44:38 <f13> the name may be different, but the action itself is the same
16:44:41 <jlaska> My preference would be to have the 2 events ... and ideally monday is quick and more of a checkpoint
16:44:55 <jlaska> is the list empty ... yes/no?
16:44:56 <f13> look at the list, make sure all the bugs are up to date with comments, seek clarification if necessary, move things around if necessary
16:45:08 <poelcat> it also gives sat and sun to fix things discussed on friday
16:45:15 <f13> jlaska: and if the list isn't empty, we have to review what's there to make sure we're no going on accurate information
16:45:31 <jlaska> f13: right on
16:45:33 * f13 plays some more pin the tail on the donkey
16:46:00 <jlaska> f13: I agree, I'd just like to continue the Friday trend and make sure that the bugs coming out of the compose have been reviewed and have data already
16:46:05 <f13> look, I don't really care what you call it, call it whatever makes sense for ya'll.  Ill just show up and do the same thing I do every other time (:
16:46:22 <f13> jlaska: that's fine, do one on Friday, but don't be surprised when we do one again on Monday
16:46:26 <poelcat> okay, i'll add it for friday, i think we can move on
16:46:32 <jlaska> yes we can
16:47:16 <jlaska> poelcat: okay ... you want  to divide and conquer for the meeting announce + summary?
16:47:24 <jlaska> or is there someone else who would like to?
16:47:55 <poelcat> jlaska: unless someone else wants to help out :)
16:48:12 <jlaska> poelcat: we can sign up adamw ... he's not here right now ... I'm sure he'd love to get action items
16:48:18 <jlaska> har har ... jk
16:48:25 <poelcat> jlaska: i'll do the recap
16:48:34 * poelcat finally has no conflicts on friday
16:48:46 <jlaska> poelcat: thx ... I'll send the announcement(s)
16:48:52 <jlaska> okay last planned topic for today ...
16:49:01 <jlaska> #topic AutoQA update from wwoods
16:49:09 <jlaska> wwoods: you've got some cool updates this past week
16:49:22 * jlaska hands over the mic
16:49:22 <wwoods> time's kind of running together, I need to figure out what I talked about last time
16:49:55 <jlaska> https://fedoraproject.org/wiki/QA/Meetings/20090727#AutoQA_update
16:50:42 <wwoods> okay, right. so - in the past week, I finished the final test
16:50:51 <wwoods> err rather - the final test of the 'installation' portion
16:50:57 <wwoods> which is to say that we now have a test script that
16:51:04 <wwoods> 1) downloads images
16:51:09 <wwoods> 2) starts a virt guest
16:51:18 <wwoods> 3) logs serial console output until loader / stage1 starts
16:51:30 <wwoods> 4) logs serial console output until stage2 gets loaded
16:51:38 * jlaska looks for something to wipe up drool
16:51:49 <wwoods> 5) downloads all anaconda logs from guest as it runs the installation
16:52:11 <wwoods> 6) declares success if anaconda finds and formats the disks
16:53:10 <wwoods> if the guest either a) goes silent for 60 seconds or more, or b) takes more than ~10 minutes without finding the disks
16:53:11 <poelcat> sounds rock'n to me :)
16:53:22 <poelcat> good job!
16:53:27 <wwoods> we declare failure and destroy the failed guest
16:53:31 <jlaska> for those interested in watching at home ... http://jlaska.fedorapeople.org/rawhide-report.png
16:53:35 <wwoods> failure is also declared if anaconda traces back
16:54:07 <jlaska> that doesn't include the install tests yet that wwoods mentioned, but gives an idea of one way that autotest can report data
16:54:28 <wwoods> it also records the kernel version, anaconda version, the complete list of critical path packages, etc. etc.
16:54:51 <wwoods> so the next obvious steps are.. make it run automatically when we get a new rawhide
16:54:57 <wwoods> and report those results somewhere useful
16:55:30 <wwoods> yeah unfortunately libvirt on ppc is... an unknown
16:55:55 <f13> where are we on the less complicated tests like repoclosure and tree sanity?
16:55:57 <jwb> wwoods, it is?
16:55:58 <jlaska> wwoods: I spoke to dcantrell last week who proposed an interesting idea for a new watcher script to help out the anaconda team ... perhaps a version 2.0 feature
16:56:11 <wwoods> f13: tree sanity is the 'sanity' portion of the test
16:56:22 <wwoods> tests 1-4 on the rawhide acceptance plan
16:56:32 <wwoods> I'm discussing the install portion - test 5-8
16:56:48 <wwoods> #link https://fedoraproject.org/wiki/QA:Rawhide_Acceptance_Test_Plan
16:57:34 <wwoods> getting public reports applies equally to both parts
16:57:47 <wwoods> but yeah, we still have tree sanity tests running
16:58:11 <jlaska> f13: that screenshot I posted above is from wwoods sanity tests
16:58:22 <wwoods> jwb: yes. I don't know anything about how/if virt works on ppc
16:58:53 <jwb> wwoods, oh.  that's easy.  libvirt could possibly work with plain qemu, but kvm and xen and lpar doesn't work at all
16:59:05 <f13> and qemu won't work for ppc
16:59:07 <jwb> wwoods, at least for any ppc platform that Fedora currently supports
16:59:10 <jwb> f13, no?
16:59:11 <jlaska> jwb: interested in doing a libvirt -> hvsi layer?
16:59:20 <f13> there aren't open firmware emulators enough to actually boot and install ppc in qemu
16:59:21 <wwoods> jwb: right, so I'll need to modify code to detect the host and modify the timeout values if we're running unaccelerated full-virt guests
16:59:50 <f13> there is no current way to do a Fedora ppc install in any Fedora provided virt
16:59:51 <jwb> f13, ah, i see what you mean.  yeah, running a ppc guest in qemu is also non-interesting to fedora
16:59:55 <wwoods> but as f13 mentions there's not a lot of value in bothering with emulated ppc
17:00:12 <f13> wwoods: it's not that there isn't value, it's that it just isn't possible to do an install
17:00:13 <wwoods> so probably the more useful direction would be to adapt the test to run on bare-metal instead
17:00:18 <jwb> jlaska, i am interested in a lot of things, but i have neither the knowledge nor the time to work on that :\
17:00:24 <jlaska> jwb: ;)
17:00:47 <wwoods> obviously that's going to require a good bit of work, but running install tests on bare metal is a desirable goal so we'll get to that. eventually.
17:01:00 <wwoods> first we need to start reporting the results of the currently written tests
17:01:29 <wwoods> and write tests for the automateable bits of the functional tests
17:01:32 <wwoods> and write a frontend to allow user submission of the remaining manual parts
17:01:52 <wwoods> checking the roadmap..
17:02:34 <wwoods> writing more tests is autoqa ticket #41-44
17:03:03 * jlaska has no clue on #41 and #42
17:03:09 <wwoods> frontend for user-submission is ticket #16-18
17:03:23 <wwoods> #41 and #42 are the X startup / X input tests
17:03:28 <wwoods> I expect those to be user-submitted
17:03:33 <jlaska> ah
17:04:06 <wwoods> I need to add tickets for "set up watcher to run current tests"
17:04:18 <wwoods> and "email results to the autoqa list"
17:04:27 <wwoods> as a stopgap measure until we get the web bits up
17:04:29 <jlaska> I think that's ... https://fedorahosted.org/autoqa/ticket/45
17:04:46 <jlaska> (for the watcher one I mean)
17:04:53 <f13> where does the simple tests like repoclosure land in the priority list?  something we could extend to not just rawhide, but updates(-testing) of previous releases too?
17:05:17 <wwoods> outside the scope of the current project
17:05:27 <f13> ugh.
17:05:34 <wwoods> the milestone is israwhidebroken.com
17:05:39 <wwoods> the tests are defined in the test plan
17:05:45 <wwoods> that being said...
17:06:39 <wwoods> ...one of the outputs of the project is a couple of test result reporting mechanisms
17:06:55 <wwoods> so wiring up something like repoclosure (which jlaska has already adapted for use under autotest)
17:07:06 <wwoods> to the test result reporting mechanism(s) we're working on
17:07:08 <wwoods> will be trivial.
17:07:20 <wwoods> so it'll kind of just happen automatically once we have test reporting wired up
17:07:49 <f13> ok.
17:07:57 <wwoods> so, if you want to help figuring out the test reporting bits
17:08:06 <wwoods> like figuring out how to get autotest sending emails
17:08:07 <f13> I'm just still harping on that point.
17:08:10 <f13> not really helping.
17:08:32 <wwoods> then we can just run the existing repoclosure test and get those emails again.
17:08:38 <wwoods> if that would be helpful.
17:08:50 <jlaska> wwoods: thanks for the update, nice work!
17:09:05 <f13> unfortunately my window for helping has closed since we're at alpha freeze and my releng duties are taking all my time :/
17:09:10 <jlaska> wwoods: any other upcoming plans to note ... or things you'd like contributor feedback on?
17:09:41 <wwoods> soon I'll probably have some docs about writing tests for autotest
17:09:58 <jlaska> delicious!
17:10:00 <wwoods> and HTML mockups (and/or mockup screenshots) of the web frontend bits
17:10:09 <wwoods> so I might ask for people to look over those things
17:10:14 <wwoods> the autotest docs are in progress already:
17:10:20 <wwoods> https://fedoraproject.org/wiki/Autotest
17:10:24 <jlaska> heh, I'm sure those are two things people will have lots of opinions on ;)
17:10:27 <wwoods> (and the associated Talk: page)
17:11:00 <jlaska> nicely done
17:11:25 <wwoods> but yeah, I'll send a message or something when those are ready for further review
17:12:01 <jlaska> okay ... I unfortunately have a conflict soon ... so only short time for open discussion
17:12:13 <jlaska> wwoods: thanks for the good news!
17:12:19 <jlaska> #topic <your topic here>
17:12:33 <jlaska> any items folks want to raise?
17:12:40 * jlaska sets the stop watch @ 5 minutes
17:16:31 <jlaska> okay ... if folks want to toss a topic, please follow-up to the fedora-test-list minutes
17:16:37 <jlaska> thanks everyone for your time
17:16:40 <jlaska> #endmeeting