/srv/irclogs.linaro.org/2012/03/09/#linaro-meeting.txt

=== liuyq is now known as liuyq-afk
plarsChiThu, zyga, springz, fabo: hi :)13:59
* fabo waves13:59
springzhello14:00
ChiThuhi14:00
* plars waits a moment to see if davepigott will hop over here14:00
fabozyga: the lock on the wiki page is gone14:01
zygafabo, what lock?14:02
fabozyga, please add your report14:02
zygaah14:02
fabothe lock that says chithu is editing the page :)14:02
plarsHi davepigott_14:03
plarsok, let's start14:03
davepigott_Hi all. Sorry for the delay14:03
plars#startmeeting14:03
linarobot`Meeting started Fri Mar  9 14:03:18 2012 UTC. The chair is plars.14:03
linarobot`Useful Commands: #action #agreed #help #info #idea #link #topic.14:03
plars#link https://wiki.linaro.org/Platform/Validation/Meetings/2012-03-0814:03
plarsfabo: do we have actions?14:03
plarsfabo: if it's easier, maybe I can change the process to capture actions in another place14:04
faboI checked the actions from previous meeting, none found14:04
plarsfabo: hmm... I really thought we did14:04
davepigott_Note: I have to leave in 40 minutes to get home before my son does. My wife still can't walk properly14:04
plarsI remember that's why I asked you about actions, because I knew we had some14:04
faboplars: I even checked the logs14:04
plarsfabo: from the one a week ago?14:04
fabowe have a couple of actions but from our sync meetings14:04
faboplars: yes14:04
plarsfabo: ok, strange but ok14:05
faboplars: http://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-01-06-02.02.log.html14:05
plars#topic Roadmap: https://wiki.linaro.org/Platform/Validation/Roadmap/2012Q114:05
faboI added this one, so everybody can see our roadmap14:06
plarsI'll give everyone a moment to check out the link14:06
plarsfabo: right, thanks for doing that!14:06
plarsfabo did a great job putting together this wiki page with our agile plan for the quarter14:06
fabofor info, it's ordered by priority14:06
plarsThere are 4 primary focus areas14:07
plarswell14:07
plarsI should let fabo explain14:07
plarsgo for it fabo :)14:07
faboheh14:07
fabo4 areas:14:07
fabolava health management14:07
fabodata policy publishing14:07
fabofast models14:08
faboand lab monitoring14:08
faboeach area have 2 sections14:08
zygafabo, fast models won't need blessing if hwpacks show up14:08
faboour agile plan (roadmap)14:09
faboand the acceptance criteria14:09
fabozyga: the blueprints evolve so it make sense14:09
fabowe're going to add another blueprint soon14:10
faboanyway, if there's something that doesn't sense to you, let us know14:10
fabo+make14:10
plarsyes, please14:10
zygaI have a question14:10
zygawe should put my work on dispatcher cloud support somewhere14:10
zygait's separate from fast models14:11
zygaand that's what I actually plan on doing14:11
plarszyga: well, it supports doing fast models though right14:11
zygaplars, it's not planned in that scope14:11
plarsjust because it can be used for a more general case, doesn't mean that it can't be applied here14:11
zygaplars, I mean, there are no work items there, no blueprints, no progress on those blueprints that exist14:11
plarszyga: it's not?14:11
zygaplars, well not here14:11
zyganot on this page14:12
fabo1st, we need a blueprint, do you have a link?14:12
zyganope, I can search14:12
plarszyga: I'm confused then, because we talked about using this idea of the dispatcher running on a remote system (cloud, or physical) to be used for fast models14:12
zygayes14:12
zygaplars, but we said that it's not really related to fast models (not waits for that, not enables that)14:12
zygaso it seems sane to put it on a separate track here14:13
fabocloud stuff are kind of generic14:13
faboI put https://blueprints.launchpad.net/lava-lab/+spec/set-up-lava-cloud-servers on fast models14:13
plarszyga: then you and I need to talk again about how the fast models piece is going to be implemented, because what I took away from our last discussion earlier this week is that we needed to run the dispatcher on the fast models host machine, so that this could be done14:13
zygafabo, that's not what Im doing14:13
fabook14:14
zygabut if you want you can keep it on that "card"14:14
zygaplars, wait14:14
zygaplars, you are getting the wrong impression14:14
zygaplars, my work will allow us to scale fast models14:14
zygaplars, if someone comes to me tomorrow with dispatcher patches for fast models that use hardware packs then great14:14
zygaplars, we could run one on control14:14
zygaplars, it's not related to what I'm doing14:14
fabo#action zyga to provide a blueprint on on dispatcher cloud support14:14
zygaplars, I'll allow any job to be sent to the cloud14:14
plarszyga: no, we don't want to run fast models on control14:15
zygaplars, I know14:15
zygaplars, bah, don't you see?14:15
zygaplars, 1) we want to run dispatcher in the cloud: check -> that's what I'm doing14:15
fabo#action plars/fabo to see where zyga's blueprint fit in the roadmap14:15
zygaplars, 2) we want to support fast models in the dispatcher: check -> it waits for hwpacks14:15
zygaplars, 1 and 2 can happen in parallel14:16
zygaplars, did you add that blueprint after our discussion14:16
zygalaunchpad is brain dead by abbreviating blueprint name to the common prefix14:16
plarszyga: so it sounded to me like you were saying 1 doesn't help with what you're currently planning for fast models - which is different from what we talked about earlier in the week14:16
zygaand I cannot really see anything https://blueprints.launchpad.net/lava-dispatcher14:16
zygaplars, it's not required for someone else (or me) to finish dispatcher support for fast models14:17
zygaplars, in that way it does not help IMO14:17
plarsso if it helps that, then it should be on that card14:17
plarsI suspect we're saying the same thing14:17
plarsand all is fine14:17
zygaI think we are14:17
plarsbut let's sync up on this later14:17
zygako14:17
zygaok14:17
fabocould we sort out that after the meeting?14:17
plarsI'm pretty exhausted right now, so maybe I'm reading it wrong :)14:17
zygasure14:17
plarsfabo: yes, that's what I just suggested14:17
zygaplars, no, we agree14:17
zyga:)14:17
plarsok, good14:18
plarsall is well then14:18
fabo#topic LAVA 2012.03 cycle14:18
fabohttps://launchpad.net/lava/+milestone/2012.0314:18
fabo#action everybody to update headline/acceptance14:19
faboplease take a look to your blueprints and make sure it's filled14:19
zygafabo, I'm removing https://blueprints.launchpad.net/lava-arm-fast-model/+spec/lava-bless-rootfs-for-fast-model from this milestone14:19
fabok14:19
faboreminder about the release date14:20
plars2 weeks from today right?14:20
fabo#info release date for LAVA 22/03/201214:20
zygafabo, what should I do with my fast model blueprints?14:20
plarswell14:20
zygafabo, they are essentially blocked until I can get hwpacks and lmc14:20
plarsfrom yesterday :)14:21
zygafabo, well I can work on one or two14:21
davepigott_fabo: FYI - We move office on the 23rd, so I will be offline most of that week14:21
plarszyga: we are supposed to have something on that soon, ricardo's team is working on it14:21
zygaok14:21
fabo#info davepigott_ - move office on the 23rd, so I will be offline most of that week14:21
plarszyga: for now, I would focus on the one to make dispatcher run on ! v.l.o14:22
fabozyga: have you discussed with riku?14:22
zygafabo, nope, I don't know that person14:22
zygaplars, ?14:22
ChiThufabo, STE asks me to work with "Include STE kernel test suite in lava-test"14:22
zygaplars, that's what the dispatcher does14:22
fabozyga: it might be a good idea to sync with him, he's the assignee from devplatform14:22
plarszyga: in the cloud, or on another machine14:22
plars! == not14:22
ubot2`Factoid 'not' not found14:22
zygafabo, thanks for telling me14:22
zygafabo, I did not know14:22
zygaplars, right14:22
fabozyga: Riku = suihkulokki14:22
plarsas in "on a system other than v.l.o"14:22
zygafabo, ok14:23
plarszyga: I think https://blueprints.launchpad.net/linaro-ubuntu/+spec/platform-enablement-fast-models is the blueprint for it14:23
plarszyga: you should probably sync with him on it, and follow it to make sure it's going to work in a sane way for us14:24
faboChiThu: once you finish your blueprint for this cycle, we can look at STE kernel test suite14:25
ChiThufabo, ok14:25
plarsfabo: do you mean the health check on?14:26
plarsfabo: I think it's done about as far as it can be right?14:26
faboChiThu: plars: I'll move this blueprint into this cycle and put it to medium for now14:26
plarsChiThu: iirc, the only one you don't have yet is vexpress, which dave is still working on, and having problems at the moment14:26
plarsfabo: +114:26
fabojsut one thing, we have 10 mins before Dave is leaving14:27
plarsah, right14:27
ChiThuplars, and MX51 which has no job since 4 months ago14:27
faboany question, status update for the lab?14:27
faboChiThu: we can remove it, we're goinf to deprecate it14:27
plarsChiThu: ignore mx51, I think we should remove the board even, it had problems that kept us from running on it that were never fixed iirc, and we moved to mx53 anyway14:28
plarssoon to be mx614:28
springzChiThu, still no commit to fix serial character truncating issue14:28
plarsChiThu: that's why mx51 wasn't on the list to do health jobs for, because we effectively have no boards there14:28
plarsright14:28
plarsdavepigott_: anything before you go?14:28
zygafabo, I'd like to request new hardware14:28
plarsdavepigott_: Did you open a bug for the ARM LT yet for the issue you are seeing?14:28
davepigott_plars: No, I think all is understood here.14:28
davepigott_plars: Yep, I opened the bug.14:29
zygafabo, my beagle is dead, I could return it and get some other boards (maybe the ones that matter more?)14:29
plarszyga: let's talk about that in aob or on email14:29
zygafabo, I've got a panda but it's very useful as a stable working system14:29
zygaok14:29
springzzyga, which type of death14:29
plarsdavepigott_: were you able to reproduce the issue by hand? doing a big extract to the sd card or similar?14:29
zygaspringz, seem my message to linaro-dev today14:29
plarsdavepigott_: and is it really corrupting the whole SD? or just leaving that partition that you are deploying in a bad state?14:30
davepigott_plars: No, it didn't fail that way.14:30
davepigott_plars: It was corrupting the boot partition.14:30
davepigott_plars: But the new build from Jan/12 from the landing team seems to have fixed that.14:30
zygafabo, updated14:30
fabothanks14:30
plarsdavepigott_: oh, so are we unblocked there now?14:31
davepigott_plars: But I can't get past the mkfs.ext3 on /testrootfs without session.run() timing out waiting for the prompt14:31
plarsdavepigott_: how long is it taking? and is that on boot or root extraction?14:31
plarss/extraction/deployment14:31
davepigott_plars: > 10 minutes - and it's on root14:32
plarsdavepigott_: I don't think 10 min. is the default timeout for that piece14:32
plarsnot for root deployment14:32
plarsdavepigott_: can you point me at a job?14:32
davepigott_plars: What is the timeout, and how do I configure it? It was on read_nonblocking() on call to mkfs14:32
plarsdavepigott_: but do it to linaro-validation in email14:32
davepigott_plars: OK.14:33
plarsno need to look at it right here, especially when you have to leave soon14:33
plarsoh wow14:33
davepigott_plars: I'll be back on it when I get home and after I've sorted out the family. :)14:33
plarsso it's not even getting to the deployment - it's still trying to format the filesystem!14:33
plarsdavepigott_: sure :)14:33
davepigott_plars: Yeah, it seems that this one is slower14:33
plarsyes14:33
plarsok, that might be pretty easy to get around then14:34
davepigott_plars: Can I increase the timeout on read_nonbloking?14:34
plarsI'm encouraged - cautiously optimistic14:34
plarsdavepigott_: essentially, you'll want to increase the timeout on the mkfs piece14:34
plarsdavepigott_: you should try it by hand and see how bad it is though, rather than trial and error14:34
davepigott_Does session.run() take an optional timeout?14:34
plarsif we are expecting it to take 10 min., and vexpress takes 2 hours or something crazy, then we need to know up front14:35
davepigott_plars: Yep. Intend to do that14:35
plarsdavepigott_: yes14:35
davepigott_plars: OK. Then I have all I need to get on with it. That's great. Feel unblocked a bit now. :)14:35
plarsdavepigott_: ok, we'll talk later today then, we'll get it sorted ou14:35
plarsout14:35
plarsgreat!14:35
davepigott_That's all from me I think. Cloud will be my focus next week and logistics of tearing down the farm and moving it.14:36
plarsdavepigott_: do we have a downtime plan yet?14:37
davepigott_plars: Intending to push one out on Monday for discussion.14:37
fabodavepigott_: is there anything we can do to get info on master images? (u-boot version, kernel, hwpack, rootfs)14:37
davepigott_fabo: I have an action to get that list published.14:37
davepigott_fabo: I have all the info, just need to pull it together14:38
plarsright14:38
plarswe discussed in our sync this week14:38
davepigott_indeed14:38
fabook14:38
plarsthanks davepigott_14:38
plarsfabo: anything else on this topic?14:38
fabonop, we can move on14:39
plarsok14:39
davepigott_OK. Bye all. See you online later14:39
fabosee you Dave14:39
plars#topic Clarify priorities14:39
plarsfabo: you have something here?14:39
fabojust a reminder14:39
faboalready discussed14:40
plarsok14:40
plarsmoving on then14:40
fabo1. reliability14:40
fabo 2. data policy14:40
fabo:)14:40
plarsright14:40
plarssee the roadmap14:40
plars#topic Clarify Yongqin blueprint assignment for this cycle14:40
faboit's a collective work to look at lava health status14:41
plarsfabo:  unfortunately he wasn't able to make it today, he told me last night14:41
fabook, skip then14:41
faboI see some discussion going14:41
plarsfabo: ok, I can give you some information here14:41
plarshe's working on part of a blueprint attached to me at the moment14:41
plarsfor doc work on lava-android-test14:41
plarsand almost done I think14:41
plarsalso he has an action to finish drafting https://blueprints.launchpad.net/lava-android-test/+spec/linaro-android-test-redesign14:42
plarswhich is in progress, and could possibly land this month14:42
plarszyga: I think you've looked at the work in progress for that change14:42
zygaplars, huh14:42
plarsdo you agree it could land this release?14:42
zygaplars, all I saw was a blueprint that's empty14:42
plarszyga: this is for the work to split out the common pieces between lava-test and lava-android-test14:43
faboplars: so https://blueprints.launchpad.net/lava-project/+spec/lava-project-doc-improvements is outdated14:43
zygaplars, I don't think we are going to see that this cycle14:43
plarszyga: I thought you had reviewed some code for it also14:43
plarsok14:43
plarsmaybe not14:43
zygaplars, not on that14:43
zygaplars, that's not started IMO14:43
zygaplars, and14:43
zygaplars, the only way it can happen in the shape of that blueprint14:43
zygaplars, is to start really working on lava-something-test that would allow us to touch bootloaders14:43
zygaplars, so hosted but non-android framework14:43
zygaotherwise it's going to be a silly redesign where we need to re-do it after we go back to bootloader tests14:44
plarszyga: yes, I think we need to talk a lot about it14:44
zygabut I still see the need to cleanup lava-android-test14:44
zygait's probably bit-roted a bit since it forked lots of old lava-test14:44
plarsno doubt14:44
zygaanyway 0.0214:44
* zyga pasted that to the blueprint14:45
zygahttps://blueprints.launchpad.net/lava-dispatcher/+spec/lava-fast-model-client <- look at the graph at the bottom14:45
zygaofftopic14:45
* zyga closing more tabs14:45
plarszyga: well, that was created by your explosion of blueprints on it :)14:46
zygaplars, yay for blueprints, blueprints are good for you, right ;-) ?14:46
plarszyga: and since doing that, we've discussed rethinking what we really need to do here, so I suspect a lot of those blueprints need to be consolidated or go away14:47
* zyga got no coding done today14:47
zygameh14:47
zygaplars, some should go away14:47
zygaI'd like to keep the separation as it is now, keeping the logical boundaries14:47
zygaand besides, less touching of blueprints is always good for anyone14:47
plarszyga: as I said yesterday, I want you, me, and fabo to have a separate meeting to make sure we're all in agreement about fast-models support14:47
zygaok14:47
plarsfabo: next?14:48
fabo#action plars/zyga/fabo to discuss about fast-models support14:48
fabo#topic progress and team issues discussion14:48
faboI've skipped some topics, I need to cleanup the template14:49
fabo#action fabo/plars to review meeting template14:49
plarsyes14:49
plarsplease :)14:49
zygaI'd like to ensure one thing14:49
zygathat after lab moves14:49
plarsI've mentioned this before... I tend to skip things that look like they are dups of other items14:49
zygawe all remember not to schedule important work for the following thee days ;-)14:49
zygaas we'll be fixing stuff on control14:50
plarsnah, it'll all just work great!14:50
plars:)14:50
zygaplars, question: any update on that lava-hw-dongle?/14:50
plars"what could possibly go wrong?"14:50
zygaplars, hehe14:50
plarszyga: no, there's a blueprint for it, it's in development14:50
zygaplars, the one you have is borked, right?14:50
plarsyes, but it was an incomplete prototype.  It's all been redesigned since then14:51
zygaplars, can we ensure that those devices have some sort of uniqueness in them?14:51
plarszyga: yes, that's the plan14:51
zygaplars, I'm pretty sure that we'll have to resort to static USB plug assignment otherwise14:52
zygaplars, and babies cry when that happens14:52
fabo:)14:52
plarszyga: I fully expect it will be a topic at the next connect, and we'll need to think about how lava needs to change, as well as our infrastructure, to support this14:52
plarson this topic14:53
plarsI'm curious about rabbitmq14:53
zygayeah?14:54
ChiThuwhat is lava-hw-dongle?14:54
plarszyga: it's deployed now?14:54
zygaplars, yes14:54
zygaplars, but on control14:54
zygaplars, I'd like to move it to the could but think it can wait a moment14:54
zygaplars, one thing I don't know about the cloud is our IP assignments and port routing issues14:54
zygaplars, so control is sane for that14:54
plarszyga: sure, but we'll need to do that for doing fast models right?14:54
zygaplars, no14:54
zygaplars, where rabbit is is irrelevant to stuff that talk to it14:55
fabospringz: Integrate tjbench to lava-test14:55
zygaChiThu, a top-secret experiment ;-)14:55
springzfabo, no input from MM team14:55
plarsfabo: I was getting to that one too14:55
springzfabo, so it is still pending, or we can make it expired?14:56
plarsspringz: they get busy and things get forgotten just like us :)  When did you ping them about it last?14:56
springzplars,  before Connect14:57
plarsspringz: oh!14:57
fabois it https://code.launchpad.net/~qzhang/lava-test/add-libjpeg-turbo ?14:57
plarsspringz: please do ask again then!14:57
springzfabo, yes14:57
springzok14:57
fabotgall_foo: ^^^14:57
* tgall_foo looks in14:57
tgall_foospringz, on lava test, have you been working with Ronynandy ?14:58
tgall_foowrt libjpeg-turbo ?14:58
plarsright14:59
springztgall_foo, I convert the input from him to this mp14:59
plarstgall_foo: I think he was just looking for feedback on it14:59
springztgall_foo, but it didn't fit for lava-test14:59
springztgall_foo, it is better to have a script14:59
zygacan I go now14:59
zygaI feel like meetings waste my time here ;-)14:59
springztgall_foo, so we're asking for a better script15:00
fabozyga: yep, we finished. except if you have blockers to talk about?15:00
zygafabo, well the board is one thing15:00
zygafabo, I really need one up15:00
tgall_foospringz, I don't know the differeence between a script / non script when it comes to lava ?      ok ?  for 12.03 rony is also going to be putting tjunittest into lava ?  sounds like you and i and rony should get together15:00
plars#topic aob15:01
plarszyga's board15:01
plarsso15:01
plarszyga: you told me about this a long time ago15:01
tgall_fooplars, could you note that as an action ?15:01
faboyep15:01
zygaplars, ?15:01
plarsand the idea was that I was going to bring you mine at the connect15:01
zygaplars, it died this morning15:01
faboeven if I'm not plars :)15:01
plarszyga: your c4 has been dead for a long time, you told me a long time ago!15:01
zygaplars, no, the SD was dead15:01
zygaplars, I bough new ones15:01
zygaplars, the board was working all the time for months15:01
plarszyga: I'm sure we can get you something though, does it have to be a c4? I don't think that's a very reasonable one to use15:02
fabo#action springz/tgall_foo/ronynandy to sync about tjunittest and tjbench work to push into lava15:02
tgall_foothanks fabo :-)15:02
plarszyga: iirc, you have an mx53 which is yours, and a panda that's linaro's right?15:02
faboyw :)15:02
zygaplars, well, if we got C4 that nobody uses that's always nice as i'm used to them15:02
plarszyga: I'd rather if you used a snowball or something that we support in the lab15:02
zygaplars, otherwise any other boards would do, I have a panda (and I'll need to de-commission it from the current task of being my arm computer to work on when needed)15:03
zygaplars, yes15:03
plarszyga: put in a request via Joey's hardware request form, and I'll approve it15:03
zygaplars, and a snowball that's dead (0-th gen)15:03
zygaplars, ok I'll find the form15:03
zygaplars, what should I choose?15:03
springztgall_foo, ok. it's not a good way to expose lots of case details in lava-test, like organize test commands and preparation in lava-test, the better way is to make a standalone case, then we can just call it to finish the test in lava-test and analyze the result15:03
zygaplars, snowball is interesting (usb serial)15:03
zygaplars, are origens available?15:03
plarszyga: I think snowball, unless you have a need for something else15:04
plarszyga: possibly, it's a bigger board, and lacks network15:04
plarszyga: you'll have to use an ethernet dongle with origen15:04
zygayeah I know15:04
plarsok, anything else? we're over time15:04
zygaplars, I'll ask on everyone, perhaps there's an unused beagle somewhere15:05
plarsok, thanks everyone15:05
plars#endmeeting15:05
linarobot`Meeting ended Fri Mar  9 15:05:40 2012 UTC.15:05
linarobot`Minutes:        http://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-03-09-14.03.html15:05
linarobot`Minutes (text): http://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-03-09-14.03.txt15:05
linarobot`Log:            http://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-03-09-14.03.log.html15:05
linarobot`Wiki:           http://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-03-09-14.03.wiki.txt15:05
fabothanks15:05
ChiThuhave a nice weekend.15:06
tgall_foospringz, sure ?  I'm not the author of the lava side of it, that's Rony.  Far as ltj goes and the test suite we have for that, I can help show you guys what is best to run and how15:06
zygageez15:06
zygait's weekend?15:06
zygaI hate weekends15:06
zygaI have to work as community15:06
zygameh15:07
zygaplars, fabo: do you want to sync now?15:07
springztgall_foo, thanks, if there is a test suite that's great15:07
springztgall_foo, I will ping Rony to get it15:07
plarsfabo: there were actions from last week!15:07
tgall_foospringz, sounds good ?  I really want to get this all done in the 12.03 timeframe15:07
plarsfabo: the log you pointed at was from january15:08
plarsfabo: http://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-03-02-02.02.log.html was last weeks meeting15:08
springztgall_foo, ok15:08
plarsfabo: it's ok though, it looks like they've all been handled already15:08
faboplars: doh, sorry15:08
fabofixing the wiki15:08
plarsfabo: the bot seems to point at the wrong location at the end of the meeting though15:09
plarshttp://irclogs.linaro.org/meeting-logs/linaro-meeting/2012/linaro-meeting.2012-03-09-14.03.log.html15:09
plarsoh15:09
plarsmaybe it just doesn't exist yet15:09
plarsthe .txt version seems to be there though15:09
plarsstrange15:09
faboplars: yeah, it's rsynced15:10
faboplars: wiki updated15:15
=== zyga is now known as zyga-afk
=== zyga-afk is now known as zyga
=== dzin is now known as dzin_lunch
=== dzin_lunch is now known as dzin
=== zyga is now known as zyga-weekend

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