IRC logs for #baserock for Thursday, 2015-09-24

*** zoli__ has quit IRC00:01
*** Walkerdine has quit IRC00:20
*** zoli__ has joined #baserock00:44
*** Walkerdine has joined #baserock00:46
WalkerdineToday I'm going to be attempting to debug the genivi platform using JTAG on the jetson00:46
WalkerdineAttempt successful thusfar01:17
*** zoli__ has quit IRC01:47
*** zoli__ has joined #baserock01:56
*** zoli__ has joined #baserock02:31
*** zoli___ has joined #baserock03:55
*** zoli__ has quit IRC03:55
*** paulw has joined #baserock06:35
*** bashrc_ has joined #baserock08:02
pedroalvarezWalkerdine: good to hear!08:05
pedroalvarezWalkerdine: you can also leave here your questions, or even use the mailing list, and we will try to help you  :)08:05
*** toscalix__ has joined #baserock08:08
*** rdale has joined #baserock08:29
*** franred has joined #baserock08:31
*** jonathanmaw has joined #baserock08:39
paulsherwoodWalkerdine: we're a tiny community bordering on a well-kept secret :/08:47
toscalix__paulsherwood: daniel and myself will meet at 10 to talk about the CIAT demo before the team sprint session, feel free to join08:49
paulsherwoodtoscalix__: would love to. can that discussion be public?08:50
paulsherwood(ie will it be here, or a hangout, or what?08:50
toscalix__it was meant to be face to face but I am hapy to bring here the conclusions08:51
*** toscalix__ is now known as toscalix08:51
straycatis anyone going to close that bracket?08:51
*** ssam2 has joined #baserock08:54
*** ChanServ sets mode: +v ssam208:54
KinnisonOh noes, now it's overclosed08:57
paulsherwooddepends on your parser08:57
* Kinnison glares in an "I have had no caffiene" way at paulsherwood 08:58
Kinnisoncaffeine too08:58
paulsherwoodso, wassup with ciat? are we there yet? are we? are we?08:58
pedroalvarezwell, that depends on your understanding of "ciat" :)08:58
* paulsherwood heard a rumour that we dealt with a systemd build breakage in an impressively short time?08:59
pedroalvarezoh yes, that was me with Kinnison's help :)08:59
pedroalvarezan easy fix but good to have it fixed before we try to move to a newer version09:00
paulsherwoodso i confess i really really dislike the buildbot view... but does what does the red in column two mean on
pedroalvarezthat the last build failed09:01
richard_mawran out of space on /src09:01
pedroalvarezI believe the disk is full09:01
paulsherwoodack. of artifacts?09:02
*** toscalix has quit IRC09:02
richard_mawhence why we made morph gc09:02
franredstraycat, could you check - I think we shouldn't block this and update in other BSPs is an orthogonal task09:02
pedroalvarezwe will need to keep this problem in mind09:02
paulsherwoodrichard_maw: ok. do you think the gc algorithm is still best/valid for the ci usecase?09:04
paulsherwoodrichard_maw: thanks for the lock feedback, btw09:04
pedroalvarezfranred: I offered my time to test it in other architectures that we think is important to have their bsp's up to date. Also there is no reason to not upgrade x86_3209:05
paulsherwoodwrt +109:05
*** toscalix__ has joined #baserock09:05
paulsherwoodbut i also think it's time we reorganised definitions09:06
franredpedroalvarez, I know, still that task could be another patch, it is completely orthogonal09:06
* tiagogomes_ deslikes that morph gc stops cleaning tempdir after X of disk space is freed09:06
rjekX is tunable I thought09:07
* pedroalvarez would like to have `morph gc --everything`09:07
richard_mawpaulsherwood: it's sufficient for continuous builders, and being able to set the minimum threshold for when artifacts are even considered for removal allows some guarantees that what you just built will survive until you deploy it09:07
rjekIf so, set it to math.inf :)09:07
tiagogomes_it is, but I still deslike it09:08
pedroalvarezfranred: well, I don't want to end up in a situation where every bsp has a different version09:08
richard_mawpedroalvarez: that's unavoidable for some platforms, but yes, I'd like if we shared the version by default09:08
paulsherwoodack, but stopping a contribution because we need more contribution is unfair to the contributor09:09
pedroalvarezpaulsherwood: fair, that's why I offered my help to review and test the patches09:09
franredpedroalvarez, IMHO, you have offered to make a different patch09:10
* straycat defers to pedroalvarez wrt 109809:10
paulsherwoodmy proposal for reorg would be scrap systems dir and have genivi, openwrt, openstack as dirs, plus a baserock dir that has buil-system, devel-system, minimal. ideally i'd put baserock default strata + component morphs under baserock dir09:12
pedroalvarezfranred: not sure about what you mean  sorry09:12
*** franred has quit IRC09:17
ssam2paulsherwood: seems reasonable in theory. So I guess you would have openwrt-specific strata under openwrt/strata/xxx, etc ?09:19
edcraggpedroalvarez: sorry, missed your ping yesterday... i'm just about to submit an updated series for partitioning (in the next hour or so), it will include documentation of the yaml format10:08
*** franred has joined #baserock10:14
franredpedroalvarez, I meant, that in my opinion, jjardon shouldn't have to rework a patch because we, as comunity, wanted to have all the BSPs at the same level. It is nice that you offered yourself to test the kernel 4.2 for different BSPs, but in my opinion this is a different patch of what we are reviewing in 1098.10:19
franredin any case is my opinion ;-)10:19
jjardonBtw, as stated in the review comments, I do not mind upgrade all the bsp as part of the patch; I only want to know what I should do to unblock the merge10:24
*** toscalix__ has quit IRC10:35
pedroalvarezedcragg: no worries, I just wanted to ask you if something was possible10:42
pedroalvarez(with your changes)10:42
pedroalvarezI believe it's not10:42
pedroalvarezedcragg: ooi, you sait that your latest patch was a WIP, what is missing from your pov to get it done?10:43
paulsherwoodssam2: yup. or maybe no need for strata subdir, depending on how many files there actually are10:43
paulsherwoodkey point is that a downstream user would create their own subdir(s) alongside genivi, openstack etc, create their own files there, refer to any of the strata/components from our upstream dirs, and never have a merge conflict for their own systems10:44
ssam2mmm, that's a good idea10:45
ssam2well, the situation for merge conflicts would be the same as now10:46
ssam2if a downstream creates strata/my-new-stratum then there won't be a merge conflict, either10:46
ssam2but still seems better than mixing everything up in one big strata/ dir10:46
edcraggpedroalvarez: it should be complete now10:46
edcraggpedroalvarez: what did you believe was not possible?10:48
ssam2one big systems/ dir, even10:48
pedroalvarezedcragg: put the entire rootfs in a non-btrfs partition10:48
paulsherwoodssam2: ack10:49
pedroalvarezKinnison, richard_maw: I believe you have some knowledge about what you think is needed to put a Baserock system in AWS. Is any of these documented? Do you have some links or a list?10:51
*** toscalix__ has joined #baserock10:51
edcraggpedroalvarez: no, not at the moment, but only because of the way the rootfs is arranged by default with subolumes, etc. that wouldn't be difficult to implement, though10:51
pedroalvarezedcragg: I agree with what radiofree said yesterday, that we will want it to have the upgradeable btrfs layout. But in some cases we will want it to be different10:52
richard_mawpedroalvarez: I had a writeup for what it *used* to require. Given the kinds of errors you're getting, it sounds like their new backing technology hasn't yet caused them to change their validation.10:52
pedroalvarezI'm temped to fake that I have grub installed and see what happens10:53
*** toscalix__ is now known as toscalix10:53
Kinnisonpedroalvarez: All I have is acquired by reading the FAWs on AWS HVM10:57
KinnisonFAQs even10:58
toscalixSummary of today's meeting:10:59
toscalix* We are on track for the CIAT demo at ELCE10:59
toscalix* We set as next check point Monday at 1210:59
toscalix* We will decide at the end of today when will it be the demo deadline, either Wed at the EOD (end of day) or Thursday at 12.10:59
toscalix* The current idea is that the last milestone should be focus in "providing a visual and comprehensive status" of the demo, so it is easy to explain and follow up the explanations about what is happening.10:59
toscalixsorry, 12:00 London time (BST)10:59
edcraggpedroalvarez: yes, it does make sense to have the option11:03
persiameeting?  Was that announced somewhere?  Which channel?11:04
pedroalvarez09:49 here (BST)11:05
pedroalvarez[09:49:23] <toscalix__> paulsherwood: daniel and myself will meet at 10 to talk about the CIAT demo before the team sprint session, feel free to join11:05
pedroalvarezit was said that it was going to be f2f11:05
persiaAh, right.  Somehow I missed that.11:06
toscalixI am open to delay the after lunch meeting 30 min. if you think it is a good idea11:15
toscalixtlsa: pedroalvarez what do you think?11:15
persiaIs that one going to be on IRC?11:15
toscalixno... f2f11:16
toscalix15 min. if everything goes as expected11:16
toscalixpersia: since I will go back home on Friday..... we will have plenty of IRC meetings the coming weeks11:17
pedroalvareztoscalix: i really don't mind :)11:18
pedroalvarezsince I moved to the UK I don't know what time is "after lunch"11:18
toscalixpedroalvarez: :-)11:21
toscalixwe keep it then11:21
* pedroalvarez tries now creating /boot/grub to see what more tests they have11:28
pedroalvarezI think having grub in baserock might not be trivial11:28
pedroalvarezheh, aws doesn't like the empty /boot/grub/grub.cfg file I've created11:58
pedroalvarezI wonder if symlinks in / to every dir in systems/default/run/* is a good idea, or just a hacky one12:15
persiaA hacky one :)  Why do you need that?12:15
pedroalvarezpersia: ask AWS :)12:16
persiaHeh.  Then just do it :)12:16
pedroalvarezpersia: I'm trying to deploy  a baserock system there, and aws complains because "the rootfs is incomplete"12:16
persiaUgh.  No obvious hits in support fora for that :(12:19
*** franred has quit IRC12:20
tlsatoscalix: I prefer the original time, as that's what I scheduled my lunch for12:21
toscalixtlsa: agree12:22
pedroalvarezheh, now i get "ServerError: an internal error has occurred during conversion"12:22
jjardonHi, can anyone else take a quick look to ?12:23
ssam2jjardon: what's the need for doing that?12:23
pedroalvarezit's sad this change will trigger a rebuild12:23
pedroalvarezbut that's why we have infra for!12:24
ssam2the idea of pre-configure-commands is that you can run something /before/ the default configure-commands, without overwriting the default configure-commands12:24
ssam2there's no harm in that change, but i don't see the point12:24
ssam2since we override the default configure-commands in this case anyway12:25
jjardonIMHO,, bootstrap or autoreconf scripts doesnt really belong to the configure step, but to a previous "bootstrap" step so the necessary files are generated to run ./configure . This change makes that more clear, related:
ssam2does that solve an actual problem?12:29
ssam2we could add an extra 'bootstrap' step, but to me it seems like extra complexity for no actual gain12:30
jjardonfor me is about correctness; one thing is to bootstrap a project so it can compile, another is to configure it with the options you want12:31
ssam2I guess. But I also think the term 'bootstrap' is overused12:32
ssam2for example, these 'bootstrap' commands won't have anything to do with GCC's bootstrap12:32
ssam2or the bootstrapping of Ruby that we do (building ruby-1.8, then ruby 2.1 using ruby 1.8)12:32
ssam2or the bootstrap branch of Vala we use (where generated C files are committed to Git)12:33
persiajjardon: Is there ever a case where one wants to perform a generic action between these two stages?12:38
WalkerdineI accidentally installed a software version of TRACE32 that was too new for my license. oops12:56
rjekSoftware licence inforcement drives people to crack software, IME12:57
persiaOnly when the code enforces it (rather than just warning).  When the enforcement is done by audit, all is good.13:03
* rjek nods13:04
*** fay_ has quit IRC13:06
*** fay_ has joined #baserock13:34
*** zoli___ has quit IRC13:51
*** zoli__ has joined #baserock13:51
pedroalvarezI don't think I'm getting anywhere soon regarding putting BR in aws13:54
richard_mawpedroalvarez: :¬(13:55
richard_mawpedroalvarez: the tooling still assuming it's got to validate everything?13:56
*** zoli__ has quit IRC13:56
pedroalvarezrichard_maw: latest error I got is "ServerError: an internal error has occurred during conversion"13:56
pedroalvarezi had the same error when I put an empty /boot/grub/grub.cfg13:57
pedroalvarez("solving" it by copying the grub folder from other OS)13:57
pedroalvarezSam thinks that aws might try to run grub itself13:57
*** Walkerdine__ has joined #baserock13:58
*** mariaderidder has joined #baserock14:01
*** zoli__ has joined #baserock14:02
*** mariaderidder has quit IRC14:06
richard_mawit at least used to use its own modified version of grub to support AWS' style of paravirtualisation14:13
*** franred has joined #baserock14:14
paulsherwoodwhat's up with the second column on the waterfall now? is it still enospace?14:24
paulsherwoodi guess i need to get onto ybd gc14:25
richard_mawpaulsherwood: it was cleared up, currently it appears to be waiting for all workers to reach the consensus that they finished14:26
richard_mawfinished as I was typing that14:26
pedroalvarezadding subcommands to ybd? /o\14:29
richard_mawpedroalvarez: not necessarily, you could do it automatically before a build if you have the right option set in your config14:30
* Kinnison sees the ciat waterfall showing a test running, having successfully built14:30
*** toscalix__ has joined #baserock14:30
*** toscalix has quit IRC14:31
paulsherwoodpedroalvarez: i wasn't intending to add commands :)14:34
paulsherwoodyou may remember i wanted morph to just do this by default... now i can make ybd do it :-)14:34
richard_mawpaulsherwood: pre-locking we wouldn't have been able to do it in morph14:36
paulsherwoodrichard_maw: oh? can you expand a bit on that please?14:36
richard_mawWe didn't want it to be automatic at the time because it would trash the staging areas left behind for debugging while they were in use14:37
richard_mawor even while building14:37
paulsherwooderk. staging and artifacts shouldn't clash?14:38
richard_mawpaulsherwood: though you may need more locking in your artifact cache to prevent it gcing something in-use14:38
richard_mawpaulsherwood: gc does both14:38
paulsherwoodah, ok14:38
richard_mawpaulsherwood: you already have auto-gc for staging areas14:38
paulsherwoodyup. now just need magic for artifacts14:39
richard_mawpaulsherwood: for auto-gc for artifacts you'll need more locking, as the "build in progress" locks won't cover it being in use14:39
richard_mawpaulsherwood: you might be able to use the same mechansim as you did for the staging areas, I don't have the time to think about it at the moment14:40
* paulsherwood was hoping just spot which files are oldest, keep touching them during builds, and assume no build takes more than (say) 48 hours14:41
richard_mawpaulsherwood: I wouldn't make that assumption of OpenOffice on ARM3214:42
richard_mawor on a Raspberry PI14:42
*** Walkerdine__ has quit IRC14:42
richard_mawLibreOffice at least allows proper parallel builds14:42
richard_mawpart of the reason LibreOffice took off so well was that the developers spent time making the tooling nice for other developers to work on14:43
*** toscalix__ is now known as toscalix14:45
*** Walkerdine__ has joined #baserock14:46
*** zoli__ has quit IRC14:53
pdarpaulsherwood: can a ybd send artifacts it has to a kbas?14:54
pdars/has/has built/14:55
*** toscalix has quit IRC14:59
*** toscalix has joined #baserock15:01
* richard_maw just broke firehose15:03
tlsaoh, meeting now?15:04
*** paulw has quit IRC15:19
Walkerdine__In which morph file is all the hardware set up for the jetson done at? There is just so many15:31
ssam2Walkterdine_: yes, they are describing a whole OS :) what do you mean by 'hardware setup' ? Linux kernel config, or other stuff?15:32
nowsterWalkerdine__: that's usually done in a bsp file and its dependents15:33
ssam2these files are probably the interesting ones:
Walkerdine__I'm just trying to wiggle a pin for now but I plan on adding spi drivers eventually so I can add a CAN chip15:34
Walkerdine__Hmm I wish it was easier to just get all the source code I need15:35
paulsherwoodpdar: yup, that's the whole idea :)15:35
paulsherwoodset kbas-url and kbas-password in ybd.conf for n instances of ybd. set password (and artifact-dir) in kbas.conf for the machine where you will run kbas15:37
paulsherwoodthe kbas machine can be a one of the ybd machines15:37
pdarpaulsherwood: cool! I have a kbas running on the aws artifact backup already and am building on my local machine15:39
paulsherwoodpdar: is it working?15:39
ssam2Walkerdine__: I guess it's just linux.git you need to modify to do that, right? the branch we use for the Jetson kernel is this one:
paulsherwoodpdar: note the passwords are sent plaintext at this point... not secure, really15:40
* paulsherwood would be delighted to accept patches for this15:40
ssam2I need to restart to test something, it will be down for 1 minute, sorry15:41
rjekCareful design must be taken for things like that15:41
rjekIt is probably easier to just use TLS, check remote identity, and continue sending passwords in the clear15:42
pdarYep, I can see it "GET" requests on the kbas but not see any artifacts that werent present being sent back15:42
pdarpaulsherwood: ^15:42
Walkerdine__What about for Genivi?15:43
jjardonHi, does baserock provide armv7hf chroot tarballs somewhere?15:44
Walkerdine__I'm trying to customize the demo platform and I need to add a whole bunch of peripherals to it15:44
*** zoli__ has joined #baserock15:45
Walkerdine__And I'm trying to keep a repositiory of all the source I need15:45
ssam2Walkerdine__: that is the kernel branch we use in any of the reference systems that are built for NVIDIA Jetson, including any GENIVI-related systems15:45
ssam2are you using Morph ? if so, the `morph get-repo` command might be useful15:45
Walkerdine__Yeah I'm using morph and I'm getting confused as to how it all works15:46
ssam2it is pretty complex to get started with...15:46
Walkerdine__I built the demo platform already on the jetson and now all I wanna do is a wiggle a pin15:47
Walkerdine__via a button on the demo platform or similar15:47
ssam2do you know the code changes that you want to make already? we can help you in this channel with getting those changes into the build tool, but i suggest #tegra if you want advice on the actual code changes you need to make15:48
Walkerdine__I should have no trouble coding it once I get the source I need15:50
*** franred has quit IRC15:50
ssam2Cool. What source do you need ?15:50
ssam2I suggest start with a clone of if you're in a Git clone of git:// -- I guess you already have that if you built the system using Morph15:50
ssam2in that clone, you can run `morph get-repo linux-jetson-tk1` to fetch linux.git and check out the ref that was built15:51
rjekpaulsherwood: Can you point me at the relevant code that sends the password in cleartext, and I'll have a look and ponder?15:51
pedroalvareznote that GDP is not merged in master, latest morph might not work15:51
ssam2same for other components, not sure what the exact names of these GENIVI components will be (I've not touched the demo platform yself)15:51
Walkerdine__Yes I checked out baserock/james/genivi-demo-platform-0.1-rebase15:51
ssam2jonathanmaw: are you working on genivi demo platform by the way? maybe you can advise better than I can15:52
Walkerdine__I also do not have the latest morph I'm pretty sure15:53
Walkerdine__Whatever comes with 15.25 I have15:53
jonathanmawI'm currently working on fixing the Yocto GDP, not the Baserokc one.15:53
* jonathanmaw reads backscroll15:53
*** petefoth has joined #baserock15:54
ssam2Walkerdine_: Morph from 15.25 should be fine15:57
ssam2jonathanmaw: ah, ok.15:57
ssam2jjardon: we don't release arm chroots as part of the release process, only the stuff in here:
*** rdale has quit IRC15:58
ssam2jjardon: you could turn the Jetson .img into a chroot though, by mounting it and copying the files out :-)15:58
jjardonok, thanks15:59
Walkerdine__So basically I just have to find the whole bunch of different parts that go into it16:00
Walkerdine__Am I still using morph to build a new image if I have all the source required16:04
Walkerdine__I'm just so used to having a makefile and the source and thats it16:05
paulsherwoodpdar: on the kbas machine, do you see the requests coming in?16:06
Walkerdine__I thought I was going to be able to have a pin wiggling over the course of a weekend. I was very wrong16:10
paulsherwoodWalkerdine__: i'm very sorry... you seem to have been on this for a couple of weeks already :/16:11
pdarpaulsherwood: I did yep, I just didn notice whether the artifacts were being pushed to the cache server if they didnt exist there16:12
paulsherwoodfwiw we're working to simplify this 'getting started' experience16:12
*** toscalix has quit IRC16:13
paulsherwoodpdar: if they are, you'll see 'uploaded' in the ybd build output, and 201 in the kbas server output16:13
*** Walkerdine__ has quit IRC16:15
* Kinnison is pleased to announce
* Kinnison will set about putting something on the :80 of it tomorrow and then announcing to the wider community16:32
pdarKinnison: Cool, what goes on :80?16:36
Kinnisonnormal http16:36
pdarI guess I should move my testing kbas to another port then?16:37
Kinnisonplesae :-)16:37
pdarpaulsherwood: Its working!16:38
richard_maw < two different landing refs being pushed, this is a separate GENIVI one, the other one also includes genivi currently16:39
Kinnisonrichard_maw: oooh16:39
* richard_maw is sad to say that you can see the set of failed firehoses when he tried to be too clever with the code16:40
*** toscalix__ has joined #baserock16:41
paulsherwoodpdar: w00t! :)16:41
jjardonKinnison: beautiful16:42
richard_mawpaulsherwood: see, someone doesn't think BuildBot is ugly16:42
Kinnisonjjardon: well, I'm hoping tlsa will make something prettier, but it represents something we've wanted for AGES for baserock, it's really exciting to have I think16:43
*** jonathanmaw has quit IRC16:43
richard_mawone of the initial goals16:43
richard_mawto be able to CI upstream16:43
* jjardon is extremely excited16:43
richard_mawgnome-continuous beat us to it :Þ16:43
Kinnisonjjardon: Once we've got the infra more stable and started on the elasticity, I'd *really* like to work with you to get a firehose running against gnome16:44
jjardonKinnison: looking forward for it :)16:44
paulsherwoodKinnison: i think we have someone itching to get stuck into that16:44
paulsherwoodnot here yet, though16:44
Kinnisonpaulsherwood: They can get a cream to keep the itching under control until we're ready :-)16:45
* Kinnison will announce all the stuffs tomorrow16:46
*** toscalix__ has quit IRC16:49
*** zoli__ has quit IRC16:56
*** ssam2 has quit IRC17:00
*** bashrc_ has quit IRC17:06
*** zoli__ has joined #baserock17:44
*** zoli__ has quit IRC18:17
*** zoli__ has joined #baserock18:22
*** straycat has quit IRC18:45
*** straycat has joined #baserock18:47
*** zoli__ has quit IRC20:09
*** zoli__ has joined #baserock20:32
*** toscalix__ has joined #baserock21:04
*** toscalix__ has quit IRC21:08
*** cyndis has quit IRC21:40
*** straycat has quit IRC21:41
*** straycat has joined #baserock21:41
*** cyndis has joined #baserock21:44
*** zoli__ has quit IRC21:58
*** zoli___ has joined #baserock21:58
*** Walkerdine has quit IRC23:57
*** Walkerdine has joined #baserock23:59

Generated by 2.15.3 by Marius Gedminas - find it at!