IRC logs for #baserock for Friday, 2014-11-07

*** zoli_ [~zoli_@linaro/zoli] has joined #baserock04:21
*** aananth [~caananth@] has joined #baserock07:00
*** aananth [~caananth@] has quit [Ping timeout: 256 seconds]08:04
*** CTtpollard [] has joined #baserock08:35
*** rdale [] has quit [Ping timeout: 260 seconds]08:37
CTtpollardif I'm not doing an devel work on a baserock VM, do I need to still attach an /src partition?08:42
paulsher1oodCTtpollard: depends what you're doing08:43
paulsher1oodthe point of having a /src partition is that you can keep your work there, and use it with various versions of baserock08:43
paulsher1oodfor example if you just do work in the normal / partition, then cycle into a new baserock version, your work isn't there08:44
petefoththe point of having a separate partition for /src is so that you can upgrade to a newer version pf Baserock without losing any files. You don’t *need* it - it’s juts a good idea to have it. What are you goin to be doing?08:44
CTtpollardthat should be fine for what I need to do08:44
petefothpaulsher1ood beat me ! 08:45
CTtpollardgetting a meteor instance and a django instance running 08:45
SotK\/home and /root are persistent between versions I thought?08:47
paulsher1oodyou may be right08:48
* persia also saw mention of force pushing in backscrol, and worries a bit08:49
persia /src is also the default location for caches, so if you need to run morph, unless you have a huge /, you will probably run out of space without /src08:49
petefothCTtpollard: CTtpollard I think that you should have a separate /src partition on your development sytem. The system youa re building and deploying may or may not need one, though I suspec that it should have one so that you can upsate the system without affecting the data08:50
*** rdale [] has joined #baserock08:53
*** ssam2 [] has joined #baserock08:57
Mode #baserock +v ssam2 by ChanServ08:57
*** wdutch [] has joined #baserock08:59
*** rdale [] has quit [Ping timeout: 255 seconds]09:02
*** mauricemoss_ [] has joined #baserock09:05
*** rdale [] has joined #baserock09:12
*** mariaderidder [] has joined #baserock09:15
*** mariaderidder [] has quit [Client Quit]09:15
*** jonathanmaw [] has joined #baserock09:32
SotKI'm writing lorries for Zuul and turbo-hipster, but I'm unsure where they should live. Should I put them under openstack-infra and stackforge respectively like they are upstream, or just in our openstack/* area?09:34
*** aananth [~caananth@] has joined #baserock09:34
*** tiagogomes [] has joined #baserock09:40
paulsher1oodopenstack/* i think09:45
paulsher1oodSotK: ^^09:45
Kinnisonprobably just under openstack/ unless you feel you really want to make another namespace09:45
KinnisonIf you do, I'd suggest infra/ and then get gerrit moved under there09:45
paulsher1oodis it possible to move lorries on gbo? i didn't realise09:45
KinnisonWell, we delete one lorry and put another one in place09:46
*** mariaderidder [] has joined #baserock09:46
Kinnisonit's not great for downstreams09:46
KinnisonOr disk space09:46
KinnisonOr cognitive load09:46
paulsher1oodso let's not do it, then :)09:46
* SotK goes with openstack/09:46
Kinnison<fx:sotto-voce>I think my plan worked</>09:47
SotKopenstack/ seemed the most sensible, I just wanted to check that no-one was opposed to it containing something not from :)09:48
ssam2only 2 VCPUs left in the DataCentered cloud09:53
ssam2it's getting quite tight in there09:53
paulsher1oodssam2: mine can be retired09:53
ssam2oh, brilliant, that frees 409:53
ssam2a few of the instances could be shrunk, but the 'resize' thing in OpenStack seems to not work09:53
ssam2the machine keeps working, but moves to 'error' state and remains exactly the same size09:54
paulsher1oodssam2: you're welcome to destroy my machine entirely if you like09:54
ssam2 we can create new devel systems quickly enough :)09:55
ssam2paul-devel is no more!09:55
SotKany chance of a quick review of this please?
pedroalvarezssam2: resizing instances will work soon! :)09:55
pedroalvarezSotK: oh! was that } overintended?09:56
paulsher1oodSotK: +109:56
pedroalvarezSotK: looks ok, +109:56
ssam2SotK: +109:56
SotKpedroalvarez: it was09:57
SotKthanks all09:57
* SotK doesn't like JSON lorries09:58
paulsher1oodpedroalvarez: would you be willing to merge my latest cycle patch, since it now fights whitespace?09:59
paulsher1oodYAML ftw09:59
pedroalvarezpaulsher1ood: yeah sure! :)09:59
* paulsher1ood realises that we now have ... very shiny :)10:00
SotKis there any reason we don't use YAML for lorries?10:00
KinnisonHysterical Raisins10:01
KinnisonIf someone sorted a patch to lorry, we'd happily switch10:01
Kinnison(well, and lorry-controller)10:01
pedroalvarezpaulsher1ood: actually I'm not very happy with it10:01
paulsher1oodwhy not?10:01
* paulsher1ood wonders if pedroalvarez means his patch, or the pastebim10:01
pedroalvarezpaulsher1ood: pastebin10:02
paulsher1oodoh - what's wrong with it?10:02
pedroalvareznot sure yet, for now I don't like that the lines are not wrapped10:03
pedroalvarezand I tried to modify the source (css) and everything went wrong10:04
pedroalvarezanyway, let's use it for now. If I find another alternative I'll discuss it here10:05
ssam2i'd prefer one that didn't require Javascript, but Storyboard and Gerrit both totally need Javascript enabled anyway10:05
ssam2so the boat has really sailed there10:05
KinnisonAvoiding JS for web applications can be hard10:06
* robtaylor likes ix nowadays10:06
robtaylortotally minimal =)10:07
Kinnisonlooks similar to sprunge.us10:10
rjekPastebins are the new photo galleries and IRC bots.  Everybody has their own.10:10
pedroalvarezssam2: we can use the fpaste implementation too10:11
locallycompactI've just tried to kill some undead lorry controller jobs on my trove by runnning lorry-controller-remove-old-jobs on said trove, is that correct? It's just hanging after five minutes.10:26
pedroalvarezlocallycompact: how many jobs are you trying to kill?10:26
locallycompactleft around after a reboot of the trove10:27
pedroalvarezah, so is a trove that already have the remove-old-jobs unit10:27
locallycompactah, it seems to have done something but the script still hasn't returned10:28
locallycompactok never mind10:29
pedroalvarezlocallycompact: what happened10:29
locallycompactpedroalvarez: The jobs all disappeared from the web view but the script never terminated in the terminal.10:30
locallycompactso I just killed it10:30
straycatlocallycompact, that zombie problem is fixed in current master10:32
locallycompactI did build from latest, maybe it was sorting itself out and I am just trigger-happy.10:33
pedroalvarezlocallycompact: anyway, you don't need to do this after further reboots, it will happen automatically10:34
locallycompact*nod*, thanks pedroalvarez straycat10:34
*** aananth [~caananth@] has quit [Ping timeout: 258 seconds]10:39
radiofreewould people be happy with having the drm driver in both the genivi and devel systems for jetson?11:04
radiofree(want to combine the bsp)11:05
ssam2as long as it doesn't break anything, sure11:05
pedroalvarezI think it makes sense11:11
radiofreehi, i broke the naming convention i think11:15
radiofreemy new branch is baserock/arm/linux-tegra-3.18-rc311:15
radiofreewhereas previously it was baserock/tegra/linux-....11:15
radiofreeshould i delete the arm/linux-tegra-* branch and push as tegra/linux-tegra?11:15
pedroalvarezradiofree: may I ask you why aren't you putting them under baserock/james?11:18
*** jjardon84 [] has quit ["undefined"]11:18
radiofreebecause i didn't assume anyone was going to review the diff between 3.17-rc5 and 3.18-rc3?11:18
radiofreei'll send some patches for review soon so you can publicly scold me there :)11:20
pedroalvarezradiofree: oh sorry, I thought I was going to see hundreds of branches for tegra :)11:20
radiofreei would like to know the procedure for this though11:20
radiofreebaserock/arm/ i think is fine, that's where most of the kernels are :\11:20
radiofreeso we have some linux-3.17+patches branch, that won't play nicely with this 3.18 branch, should i first push baserock/james/linux-tegra-3.18-rc3 first?11:21
radiofreethey get authorisation to push it as baserock/arm/linux-tegra-3.18-rc3?11:21
pedroalvarezradiofree: IMO, I think we should use personal branches (baserock/pedroalvarez/* in my case) when something may not end up in definitions11:23
pedroalvarezbut is my opinion11:23
pedroalvarezI think there isn't anything written down about this11:23
radiofreepedroalvarez: i'm planning to send patches for definitions11:23
radiofreeupgrade kernel is one of them11:23
pedroalvarezI know, (and I looking forward to test that patch)11:24
pedroalvarezregarding putting them under baserock/arm, or baserock/tegra, we are aiming to use the same branch for all of them11:25
radiofreewell this 3.18-rc3 kernel should be fine on most of our arm platforms11:25
radiofreeit does, however, contain a few jetson specific patches that are not mainline yet11:25
pedroalvarezand I understand that11:26
pedroalvarezradiofree: so regarding the naming convention, it's your call11:27
pedroalvarezyou can remove the branch that you have just created since nobody has ever used it11:28
pedroalvarezand because we are 100% is not being used in our definitions11:28
pedroalvarezand if anybody disagrees with my opinion, please shout11:29
radiofreei prefer the current name tbh, since any arm system should be able to use it11:34
pedroalvarezdo yo think that the jetson specific patches would affect other architectures? (x86, ppc..)11:37
radiofreeso yeah, it's a pretty generic 3.18-rc3 kernel that any system could use11:40
radiofreethought for that it might be better to wait for 3.18 proper11:40
radiofreeso i'll use the baserock/jetson/foo convention until that can happen?11:41
radiofreepaulsher1ood: btw i built a baserock system on the chromebook yesterday11:41
* richard_maw is looking forward to the Amazon fairy bringing him his own11:45
paulsher1oodrichard_maw: there's one waiting for you11:46
* richard_maw is confused11:48
richard_mawI was referring to one I placed an order for11:48
paulsher1oodwell i'll give yours to someone else :)11:48
radiofreeis that the one simon is currently taking apart?11:49
pedroalvarezoff topic!11:50
*** cosm [] has quit [Ping timeout: 245 seconds]12:45
jmacsDid someone say they'd had apt running on a baserock linux system? Or did I imagine that?13:01
straycatWe have pip, I don't know of anyone running apt.13:02
*** abdul [~abdul@] has quit [Quit: Leaving]13:02
pedroalvarezI can't see nothing about that in the irc logs13:04
jmacsI probably imagined it.13:04
radiofreewhere does the baserock-jetson-flash script live?13:11
richard_mawI have no idea, I though you were the one who made it.13:14
franredjmacs, I think someone said that persia has some branch where he is able to install debian packages, but I've never seen it and also Im not sure if Im right13:20
jmacsWell, I've built dpkg, so I'm going to give it a try13:22
rjekSo, who wants to jail-break and port Baserock to the Amazon TV?13:29
Kinnisonpersia wrote a tool to munge debs onto a BR system for bootstrapping iirc13:30
radiofreerichard_maw: yeah, i did, but how does it get included with the jetson releases these days?13:31
radiofreei manually made the archive13:31
richard_mawAFAIK we've not done anything special, and the process involves taking one of your old release tarballs and replacing the images included. I'd like to be proved wrong.13:34
radiofreeok, well i'll make a change to it and send it to the list for review, then hopefully you can tell me how to make it less crap13:36
*** zoli__ [] has joined #baserock13:42
*** zoli_ [~zoli_@linaro/zoli] has quit [Read error: Connection reset by peer]13:42
straycat is now known as vletrmx2113:54
pedroalvarezi did the last release, and I manually put the script and the images together14:06
pedroalvarezI didn't have the time to think about how should it be done, and to automate it14:07
paulsher1oodhow would folks feel about doing a genivi release next week?14:08
pedroalvarezlet me check the changes14:09
vletrmx21 is now known as straycat14:11
*** cosm [] has joined #baserock14:17
radiofreepaulsher1ood: when?14:19
radiofreei'd like to get a few things in there14:19
pedroalvarez14th nov14:20
straycatjmacs, why do you want apt-get anyway?14:38
jmacsBecause our client's script uses it.14:39
franredcan you not rewrite the script and add the packages before it uses any kind of package installer?14:40
straycatrelying on any package manager kind of misses the point of using baserock14:40
jmacsOK, I'll just tell our client that.14:41
franredjmacs, we add the packages in the baserock way usually and tell the clients how we have done it, me though14:42
jmacsRight, I can go into more detail now that meeting's over14:47
jmacsFirstly, there are hundreds of packages to install14:47
straycatwhat meeting?14:48
jmacsSecondly, converting their script to use baserock exclusively would fork it14:48
jmacsand I'd have to update it for all their changes14:48
jmacsThirdly I don't have any good reason at all to believe you can separate out the software installation and configuration steps which are currently interleaved14:49
persiaMy tool was indeed something to munge debian packages into arbitrary binary-compatible systems.14:50
jmacsIf apt/dpkg could be made to work on baserock it would solve a lot of problems quickly (although quite probbably not in an ultimately correct way)14:50
jmacspersia: That may well be useful.14:50
persiaThere are some outstanding bugs, but looking at the fpc branch is probably most useful.14:51
franredapt and dpkg are not baserock tools as far as I know. Im facing the same problem as you in openstack, adding multiple python packages to the strata and have to configure some of them14:51
ssam2jmacs: I'm sure you can get apt-get running in Baserock, I did it with Yum once as an experiment. The problem is that the first thing you install will try to pull in the whole of Debian as dependecies14:51
jmacspersia: Thank you.14:51
franrednot sure what is the value to use baserock if you are going to use dkpg or apt-get as ripsum said before14:51
persiaFor an ugly solution to package configuration, take a look at how fpc was done in Baserock.14:51
ssam2jmacs: so you basically end up overwriting Baserock with Debian, or creating a Debian chroot inside the Baserock system14:52
persiaA better solution is integration scripts for the build environment, but we never finished those.14:52
jmacsWhat does an integration script do?14:53
persiasystem integration scripts run at system assembly time, to do thinks like initialise the runtime link mappings, etc.14:54
persiabuild integration scripts are not currently supported by morph.14:54
* persia is vaguely unhappy with putting something not from in the openstack/ lorry namespace, but doesn't have a better suggestion14:57
*** petefoth_ [] has joined #baserock14:58
*** petefoth [] has quit [Ping timeout: 260 seconds]15:01
petefoth_ is now known as petefoth15:01
SotKpersia: whats that in reference to?15:28
jjardonradiofree: is to fix this build failure:
jjardonwe can appply this less intrusive patch as well :
richard_mawyou can override some make variables from the command line, so it's possible we may not need to patch the makefile15:46
richard_mawbut I don't know what variable that would be post-automake-expansion15:46
radiofreejjardon: is this because of upgrading glibc?15:48
radiofreein my 14.40.1 system i compile 1.6.0 all the time and don't have a problem15:48
richard_mawjjardon: I think you could get away with changing the make command to `make weston_multi_resource_LDADD='$(SIMPLE_CLIENT_LIBS) -lrt -lm'`15:55
*** zoli__ [] has quit [Remote host closed the connection]15:58
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock15:59
jjardonrichard_maw: thats what we are doing now, but I though the proper fix is to patch the Makefile.am16:05
richard_mawpatch and submit upstream, until then we prefer to minimise our delta16:06
jjardonradiofree: no, I do not think so but the fact that we have what richard_maw suggest in our morph file suggest this problem was hitted by someone else before16:06
pedroalvarezjjardon: s/hitted/hit/ :)16:07
jjardonrichard_maw: its already submitted ;) the patch seems to be correct but a check in is missing16:07
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]16:08
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock16:09
* paulsher1ood hopes his cycle patch will land soon :)16:10
pedroalvarezpaulsher1ood: ooopppsss16:10
pedroalvarezI've been busy16:10
*** wdutch_ [] has joined #baserock16:17
*** wdutch_ [] has quit [Client Quit]16:17
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]16:18
*** zoli_ [] has joined #baserock16:18
*** zoli_ [] has quit [Changing host]16:19
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock16:19
persiaSotK: Earlier discussion about how to lorry something (I didn't quite catch all the context).  My thought is that we want to avoid using namespacing where things might end up used in other contexts, but I don't feel strongly enough to suggest undoing anything done, or reopening discussions complete.16:21
straycatHrm, seems that python requirements are actually ambiguous16:22
persiafranred: One reason to use packages from other places is for self-hosted toolchains.  Many of these are hard, or even impossible to bootstrap, but by using a binary blob from somewhere, one can often build things, which then stay built.  We need to think more about how to ensure this doesn't need to be repeated, but I don't think we reached consensus on that conversation yet.16:23
persiastraycat: They can be: the wheel format is supposed to address some of that, but not everything is wheels yet.16:23
SotKpersia: oh, those things were on, just not in the openstack/ prefix on there like everything else in openstack/ is on git.baserock.org16:24
persiaHmm.  I don't actually know what that means: I'll ask.16:25
franredpersia, I agree if there are no other way to boostrap something we can add it, but not as a general rule. OMHO16:25
persiafranred: Absolutely.  Otherwise there is no point.  As soon as one runs dpkg or rpm, the system becomes an unrecoverable snowflake.16:26
straycatpersia, what i've just found is that foo<1.6,>1.9,==2.4c1 can mean be interpreted as both valid and invalid16:26
persiaheh :)16:27
franredstraycat, from a human perspective I would always take the latest version so ==2.4c116:27
straycatit's a known problem and treating it as invalid is on the todo list fortunately16:27
persiafranred: The latest may not be API compatible with the code expressing the requirement16:28
straycatfranred, it's an invalid requirement really16:28
paulsher1oodpedroalvarez: thanks!16:31
ssam2straycat: maybe you should postpone working on validation if it's turning out to be complicated16:32
straycatssam2, it's not too bad16:33
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]16:50
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock16:50
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]16:51
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock16:52
* jjardon just discovered the -enable-kvm option of qemu and now everything is FAAASSST :)16:52
jjardonso I have a system running with systemd 217 here. Anything I should test before sending the patches?16:57
paulsher1oodooh, exciting :)16:57
*** CTtpollard [] has quit [Quit: Ex-Chat]16:58
ssam2jjardon: if networking works and 'systemctl list-units' doesn't show anything unexpectedly broken then, cool17:05
*** cosm [] has quit [Ping timeout: 260 seconds]17:16
*** thecorconian [] has joined #baserock18:03
*** thecorconian [] has quit []18:03
*** flatmush [] has quit [Ping timeout: 240 seconds]18:04
ssam2good progress with the Baserock OpenID provider today, all I need to do is work out how to deploy it with a real webserver and we'll have a usable prototype, I think18:08
ssam2less progress with Storyboard, it turns out the Puppet scripts for deployment provided by the OpenStack project on Github don't work out of the box18:09
*** flatmush [] has joined #baserock18:15
*** wdutch [] has quit [Quit: Quit]18:15
*** jonathanmaw [] has quit [Quit: Leaving]18:16
*** mauricemoss_ [] has quit [Quit: Leaving]18:21
*** mariaderidder [] has quit [Quit: Ex-Chat]18:24
*** tiagogomes [] has quit [Ping timeout: 244 seconds]18:25
*** ssam2 [] has quit [Quit: Leaving]18:26
jjardonwhat parameters should I pass to my qemu to have internet connection inside the vm?18:28
*** locallycompact [] has quit [Ping timeout: 265 seconds]18:30
*** franred [] has quit [Quit: Leaving]18:31
*** sambishop [] has quit [Remote host closed the connection]18:37
robtaylorjjardon: thats a deep and tricky question. Natted? 19:38
juergbithe default should just work for simple internet access from the vm19:41
*** vmeson [~quassel@] has quit [Ping timeout: 244 seconds]19:42
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]19:56
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock19:58
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]20:03
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock20:04
*** cosm [] has joined #baserock20:40
*** rdale [] has quit [Ping timeout: 244 seconds]21:02
*** vmesons [~quassel@] has joined #baserock21:09
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]21:19
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock21:19
*** zoli_ [~zoli_@linaro/zoli] has quit [Remote host closed the connection]21:26
*** zoli_ [~zoli_@linaro/zoli] has joined #baserock21:26

Generated by 2.15.3 by Marius Gedminas - find it at!