IRC logs for #baserock for Tuesday, 2015-02-17

*** zoli___ has joined #baserock03:59
*** zoli__ has quit IRC03:59
*** gfinney has joined #baserock08:07
*** gfinney_ has joined #baserock08:07
*** a1exhughe5 has joined #baserock08:11
*** a1exhughe5 has quit IRC08:19
*** rdale_ has joined #baserock08:22
*** gfinney has quit IRC08:23
*** zoli___ has quit IRC08:23
*** gfinney_ has left #baserock08:23
*** gfinney has joined #baserock08:24
*** zoli__ has joined #baserock08:46
*** mariaderidder has joined #baserock08:49
*** mike has joined #baserock08:53
*** mike is now known as Guest9363608:54
*** mauricemoss_ has quit IRC08:55
*** mauricemoss_ has joined #baserock08:55
*** bashrc has joined #baserock09:05
*** jonathanmaw has joined #baserock09:21
*** tiagogomes_ has joined #baserock09:24
*** gary_perkins has joined #baserock09:30
*** ssam2 has joined #baserock09:41
*** ChanServ sets mode: +v ssam209:41
*** lachlanmackenzie has joined #baserock10:07
*** lachlanmackenzie has quit IRC10:12
*** edcragg has joined #baserock10:17
*** Krin has joined #baserock10:17
*** wschaller has joined #baserock10:25
*** lachlanmackenzie has joined #baserock10:32
edcraggwould it still be acceptable to use the nfsboot.write extension for deployment to a netboot server in the HP moonshot, even though the source says it's depreciated? It seems to do exactly what's needed in this case, i.e. putting a baserock rootfs and kernel on an existing netboot server, where the pxeboot.write extension does more things which aren't required10:34
pedroalvarezedcragg: hmm.. deprecated? do we say that somewhere?10:37
edcraggpedroalvarez: yes, in a comment at the top of the nfsboot.write script10:39
edcragg*** DO NOT USE ***10:39
edcragg- This was written before 'proper' deployment mechanisms were in place10:39
tiagogomes_there is a nfsboot-server.configure write extension10:41
tiagogomes_on definitions10:41
pedroalvarezI guess he wants a deployment extension, no a configuration exts10:41
tiagogomes_ah right10:41
richard_mawnfsboot-server.configure  is some of the work to set up nfs booting that works for nfsboot.write10:42
edcraggthat doesn't do anything useful iirc in this case10:42
richard_mawbut you also need to fiddle with the bootloader on each node for nfsboot.write to work as expected10:42
edcraggrichard_maw: how so? The moonshot U-Boot's already set up for network booting by default, all we need is to deploy new baserock systems to the netboot server10:44
pedroalvarezedcragg: I'd say: use nfsboot.write server :)10:47
edcraggpedroalvarez: ok, cool, thanks :)10:48
pedroalvarezedcragg: althoug maybe you can do the same thing with pxeboot :)10:48
richard_mawedcragg: the nodes need to be set up to know "who" they are, and it needs to be synchronised with what you specify when you deploy with nfsboot.write10:54
edcraggpedroalvarez: i did wonder that, it seems to set up a local nfs/tftp server and remotely boot an installer system from that if i've understod correctly, which is sort of the other way around10:55
richard_mawpedroalvarez: using pxeboot depends on if you want it to boot from the hard-drive, or always boot from PXE10:56
richard_mawedcragg: broadly, yes, though I believe it can be configured to use an existing tftp/nfs server instead of spawning one10:56
edcraggrichard_maw: ok, thanks, i'll have a look10:57
edcraggalso, understood about having some sort of device ID10:58
edcraggi was thinking this could be possible with a MAC specific pxe configuration on the server possibly10:59
richard_mawyeah, you ought to be able to use the MAC address as the unique name11:00
richard_mawit'll save having different config on each node11:01
edcraggyep, it could be the only option anyway as we have no real way to configure U-Boot11:02
richard_mawyou can't save changes to your u-boot config?11:02
* richard_maw sighs11:02
richard_mawit was the same for the Highbank servers11:02
richard_mawwe decided instead to have it use a u-boot script it read from the HDD, which just told it to NFS boot11:03
* richard_maw has looked at OpenStack's Ironic service for bare-metal deployments recently.11:05
richard_mawIt would be a nicer way to do this, if it weren't for the fact that there's a huge amount of work involved in getting it set up, plus extra engineering to put together a suitable installer image for moonshot nodes11:06
tlsaduplicate lorries?11:11
pedroalvareztlsa: looks like that11:11
tlsaone used in strata/libdrm-common.morph11:12
tlsathe other used in strata/virtualization.morph11:12
pedroalvareztlsa: thanks for the info!11:14
pedroalvarezgiven that libdrm-common is part of genivi, and virtualization has never been in a release, I'd vote for removing the one used in virtualization.morph. But that would break the definitions before the change11:16
tlsawell, I noticed because x complains that can't be found when it starts on my latest xfce system11:17
tlsaso, I was assuming that x needs it11:18
radiofreeyou'd want to use libdrm-common for x as well11:20
radiofreedepending on the type of x driver you want to build11:20
radiofreewe could try adding to the jetson image11:22
radiofreetlsa: does it work? or fail because of that?11:24
tlsait doesn't work, but its unclear if that's the cause.11:25
tlsait reports that that lib isn't found, then sits there for a while, then says it failed to start11:25
radiofreewhat does the xorg log say?11:26
radiofreeif you're running in a vm you'll probably need
tlsaradiofree: rebooted into the dev system11:26
tlsaradiofree: I've got the fbdev driver (was confused for a moment because git grep didn't show it, but that was because git doesn't know about my strata/x-xorg.morph yet)11:32
flatmushHey, does anybody know if I can run the baserock license checking scripts over a set of repositories managed in repo, and if so, where would I find this script?11:34
radiofreetlsa: could you paste the Xorg.0.log somewhere?11:34
pedroalvarezflatmush: (licenseckeck.(sh|pl)11:36
flatmushpedroalvarez: Thanks11:36
radiofreetlsa: you need to add vga=somemagicnumber to your kernel args11:39
tlsawhere do I do that?11:39
radiofreevga=791 for now (1024x768x24)11:39
radiofreevga=791 for now (1024x768x16) even11:39
radiofreetlsa: `mount /dev/yourrootpartition /mnt`11:40
radiofree`vim /mnt/extlinux.conf`11:40
franredpedroalvarez, the default layout in config.layout will put the things in /usr/local/apache2, is there where you suggest to install httpd-apache libraries/binaries...?
radiofreeedit the kernel args for your TEST, add vga=791 at the end11:40
pedroalvarezfranred: I see more than one layout there11:42
radiofreetlsa: `mount /dev/sda /mnt` for you11:42
franredpedroalvarez, umm, I may missread your review, I though you suggest the default layout (I think to use the RedHat layout instead)11:42
franredsorry for the noise and thanks for the review :)11:43
franredpedroalvarez, ^^11:43
*** tiagogomes_ has left #baserock11:44
*** tiagogomes_ has joined #baserock11:44
pedroalvarezfranred: Redhat layout hardcodes /usr :(11:44
pedroalvarezbut is better than the previous layout :)11:45
pedroalvarezactually all of them harcode /usr11:46
pedroalvarezfranred: I'd suggest something like: ./configure --prefix="$PREFIX" --enable-layout=RedHat ............11:47
tlsaradiofree: now xinitrc:85: xrdb: not found, and xfce4-session says not found11:47
tlsaI'll look into it later - probably more missing deps11:48
franredpedroalvarez, I've already had that... although I override some of the layou configuration :)11:48
pedroalvarezfranred: true, then don't override anything else then, I guess11:50
*** rdale has joined #baserock11:50
franredpedroalvarez, I will give it a go. thanks11:50
radiofreetlsa: well you're getting somewhere at least11:51
radiofreetlsa: does it start normally without using xfce?11:52
*** rdale_ has quit IRC11:52
*** rdale_ has joined #baserock11:53
*** rdale has quit IRC11:56
tlsaradiofree: yep, it does11:58
radiofreeat least x is working then!11:59
tlsawhat's the correct way to set the kernel args?12:00
tlsain the cluster morph?12:00
tlsausing the upgrade-devel.morph one atm12:01
radiofreeyou only need to specify the additional ones, so in your case KERNEL_ARGS: vga=79112:02
*** gfinney_ has joined #baserock12:24
*** gfinney has quit IRC12:24
jjardontlsa: add gtk-deps stratum to your system definition12:38
jjardontlsa: libdrm-common and graphics-common as well12:50
*** mariaderidder has quit IRC12:52
*** mariaderidder has joined #baserock13:20
*** Blacksnow has joined #baserock13:23
*** Krin has quit IRC13:23
*** Blacksnow has quit IRC13:23
*** Krin has joined #baserock13:23
*** fay_ has quit IRC13:29
tlsajjardon: yep got them13:59
*** fay_ has joined #baserock14:00
*** rdale_ has quit IRC14:02
*** rdale has joined #baserock14:02
*** Krin has quit IRC14:06
rdalethe baserock image is was developing with isn't working properly anymore. it has 'factory' and 'TEST' on it, and 'TEST' still boots fine. is it possible to boot under 'TEST' and then run btrfsck on the 'factory' partition?14:32
pedroalvarezrdale: do you want to recover the hole factory subvolume? or only some files/folders?14:33
rdalei would like to recover the whole thing if possible14:33
rdalemaybe it would be easier to just start again - my /src partition seems ok where i have my current work14:34
radiofreemason failing to connect to gbo14:43
radiofreeoh, maybe it's fine?14:43
radiofreethere's nothing new to build i suppose14:43
pedroalvarezyeah, nothing new to build. Is good to know when g.b.o has had network problems15:01
*** rdale has quit IRC15:04
*** wschaller_ has joined #baserock15:08
*** Codethink__ has joined #baserock15:10
*** rdale has joined #baserock15:11
*** wschaller has quit IRC15:11
*** wschaller_ has quit IRC15:13
*** zoli__ has quit IRC15:29
radiofreefatal: unable to access '': Peer's certificate issuer has been marked as not trusted by the user.15:42
radiofreedid i do that?15:42
radiofreei mean did i mark it as not trusted?15:42
ssam2maybe it's automatically not trusted because it doesn't exist, or is some random self-signed certificate15:43
ssam2I don't know what if any certificate it's using, but we definitely don't have one in place that's signed by a root certificate authority that you'd be likely to have in your system's default trusted roots file...15:44
persiardale: When I've broken my environment, I've generally just deployed a new devel image somewhere, and reattached my /src, with little impact.15:45
persiaThe main pain point is if morph changes require definitions changes, which means I have to rebase when I get a new devel system, but I try to rebase every day or two anyway, so this isn't onerous.15:45
rdaleyes, it isn't too bad. it was just that i was in the middle of making some good progress and didn't fancy spending an hour or two creating a new image15:46
*** jonathanmaw has quit IRC15:48
*** jonathanmaw has joined #baserock15:48
paulsherwoodrdale: there is nothing special about factory, now. you could adopt TEST as your default working state15:49
paulsherwood(and remove factory)15:49
* paulsherwood these days names his working volume based on the date of creation eg FEB15 is my current system)15:50
rdaleyes, i'm using to create an environment called 'development' from a vanilla build image15:50
radiofreejjardon: how do you build xterm?15:50
paulsherwoodrdale: :)15:50
jjardonradiofree: you need libXaw, which pull quite a few more dependencies; check this commit to know what you need:15:54
* radiofree won't bother15:56
radiofreehmm actually it's not that many15:57
radiofreexwayland works on the jetson16:09
* pedroalvarez noticed today an /usr/ebin folder in a system16:18
pedroalvarezI thought it could be a typo somewhere, but looks like it something usual16:19
jjardonradiofree: nice!16:28
*** zoli__ has joined #baserock16:50
*** mariaderidder has quit IRC16:55
* richard_maw doesn't even like sbin17:00
richard_mawwhat is ebin for?17:00
richard_mawerlang programs?17:00
persiaI thought it was for people whose left hand had limited independence of finger movement.17:00
rjeksbin's nice, because it removes binaries that normal users have no use for out of the tab completion list.17:01
* rjek mutters at debian putting ifconfig in /sbin17:01
pedroalvarezrabbitmq-server is intalling this folder17:02
* pedroalvarez mutters at busybox installing things in /bin /sbin /usr/bin and /usr/sbin17:04
paulsherwoodpedroalvarez: we need to try toybox17:05
pedroalvarezI ended up removing /bin and /sbin of my devel system building busybox..17:07
pedroalvarezI don't know what happened, maybe it was just me doing things wrong, but I normaly know what I'm doing.. :/17:08
* pedroalvarez will investigate17:08
jjardonpaulsherwood: what are the advantages of toybox over busybox?17:10
pedroalvarezjjardon: note that he didn't say "replace busybox with toybox" :)17:11
pedroalvarezWould be nice to have alternatives. IIRC someone was looking at musl as well17:12
tiagogomes_what are the advantages of toybox and busybox when we keep replacing its functionality? like recently for the diff command17:14
* paulsherwood believes that toybox ultimately replaces busybox and will do a better job17:15
persiaFor smaller systems, I think toybox is newer and more loosely licensed.  For larger systems, I think we should use GNU, and neither of busybox or toybox.17:16
persiaAnd I think it's really important that we have both examples in the reference systems, so violently disagree with any claim that "Baserock is replacing busybox", despite any changes that may be landing to the contrary (just because not enough of us care about little environments).17:17
paulsherwoodpersia: yup, except that i think toybox has a reasonable chance of superseding gnu for some classes of larger systems too17:17
persiaI disagree, mostly because GNU is so widespread on larger systems, and so many high-level components depend on GNU extensions to build or run.17:18
persiaBut over time, I may be proven wrong :)17:18
paulsherwoodyes, me too :)17:18
tiagogomes_if the toybox tools don't use the GNU extensions, it is not going to be easy to replace with toybox17:19
richard_mawif toybox doesn't get held back too much by micro-optimisations to code size it could get there, there's a lot of cruft in GNU coreutils that's there to deal with having to support multiple platforms17:20
paulsherwoodtiagogomes_: perhaps. but the toybox maintainer explicitly aims to make it a viable alternative to GNU in some contexts17:21
tiagogomes_v8 doesn't have a install target ¬¬, it seems designed to be embedded in the source of the software that makes use of it17:22
ssam2the Toybox maintainer also produces a distro similar to Baserock (but smaller in scope):
ssam2so Toybox will probably support most of what Baserock needs to build, eventually :)17:24
pedroalvarezHas anybody hit this error when trying to mount a baserock.img disk?
pedroalvarezIs odd, this Debian can mount other baserock VMs but not the one that we are trying to mount.17:36
pedroalvarezAnd as far as I am aware, we haven't changed btrfs currently17:36
*** zoli__ has quit IRC17:41
radiofreewas that an image build from a build-system built from master?17:44
pedroalvarezradiofree: your question scares me17:45
pedroalvarezradiofree: I think so17:45
tiagogomes_I updated the btrfs userland recently for aarch6417:45
radiofreeimage built from a build-system from master17:45
radiofreebtrfs-progs was update 5 days ago17:45
tiagogomes_maybe running the command with strace will help...17:45
radiofreeand we had a kernel upgrade merged a few days ago17:45
pedroalvarezthe problem is just Debian, that don't support that btrfs version17:46
pedroalvarezit can be mounted inside of baserock17:46
radiofreewhich version of debian?17:46
pedroalvarezpdar: which version of debian are you running?17:47
radiofreeif this is your laptop, and it's wheezy, upgrade17:47
pdarpedroalvarez: radiofree: wheezy 7.817:48
radiofreei've no idea how to upgrade debian17:48
pedroalvarezhaha, I think he is running latest17:49
radiofreei think you can upgrade to "jessie"17:49
pdaryou can17:50
pdarI shall17:50
pedroalvarezpdar: try `apt-get dist-upgrade` first17:50
pdarpedroalvarez: nothing changed17:51
radiofreei don't really know what "Debian Edu" is though17:52
pdarno, I found something similar somewhere else though, so shall proceed17:53
pdarthanks radiofree17:53
*** bashrc has quit IRC17:57
*** tiagogomes_ has quit IRC18:02
*** zoli__ has joined #baserock18:02
*** jonathanmaw has quit IRC18:02
*** lachlanmackenzie has quit IRC18:06
*** zoli__ has quit IRC18:12
*** edcragg has quit IRC18:28
*** pedroalvarez has quit IRC18:43
*** pedroalvarez has joined #baserock18:44
*** ChanServ sets mode: +v pedroalvarez18:44
*** ChanServ sets mode: +o pedroalvarez18:45
*** zoli__ has joined #baserock18:47
*** Codethink__ has quit IRC18:47
*** zoli__ has quit IRC18:51
*** rdale has quit IRC18:53
*** zoli__ has joined #baserock18:53
*** Guest93636 has quit IRC18:57
*** ssam2 has quit IRC19:12
*** gfinney_ has quit IRC19:13
*** gfinney__ has joined #baserock19:13
*** gary_perkins has quit IRC19:42
*** gfinney__ has quit IRC19:47
*** gfinney_ has joined #baserock19:47
*** zoli__ has quit IRC20:46
*** elevenarms has joined #baserock20:46
*** gfinney_ has quit IRC20:47
*** gfinney__ has joined #baserock20:47
*** zoli__ has joined #baserock21:05
*** JPohlman1 is now known as JPohlmann21:38
*** JPohlmann has joined #baserock21:38
*** gfinney__ has quit IRC21:47
*** gfinney_ has joined #baserock21:47
*** elevenarms has quit IRC22:04
*** elevenarms has joined #baserock22:05
*** elevenarms has quit IRC22:09
*** elevenarms has joined #baserock22:16
*** gfinney_ has quit IRC22:48
*** gfinney_ has joined #baserock22:48
*** zoli__ has quit IRC23:27
*** gfinney__ has joined #baserock23:48
*** gfinney_ has quit IRC23:48

Generated by 2.15.3 by Marius Gedminas - find it at!