[07:00] good morning [07:11] good morning === erkules_ is now known as erkules [07:44] Morning all o/ [08:41] mo'in [08:42] moin moin [08:47] hey Chipaca [08:47] mvo: hi there [08:47] mvo: how's the sprint? what have you broken *now*? [08:48] :) [08:49] Chipaca: sprint is interessting, we can talk next week :) [08:49] mvo: ok [08:50] mvo: wink your left eye if you are held against your will and can't talk about it [08:50] lol [08:51] * mvo winks [08:51] lol [08:51] * Chipaca hopes that was the right eye [08:53] wasnt the left eye the right eye ? [09:07] ogra_: only if was held against his will === vrruiz_ is now known as rvr [09:24] Good morning all; happy Friday, and happy Chocolate Pudding Day! 😃 [09:25] yummy !! [09:25] though i prefer icecream at this weather [10:06] ogra_: mmmm, ice cream! [10:09] :D [10:10] (sadly i cant ... recovering from a jaw infection kind of goes against it :/ ) [11:12] * Chipaca -> lunch [11:23] sergiusens: fwiw, I triggered a build with the new livecd rootfs so that we can test the deb/apt less system [11:24] well, semi-deb-less :) [11:24] dpkg itself is still there [11:25] I have just installed snappy in beaggle board, following steps in https://developer.ubuntu.com/en/snappy/start/#try-beaglebone [11:25] ssh ubuntu@webdm.local [11:25] ssh: Could not resolve hostname webdm.local: Name or service not known [11:25] sergiusens, the shadow patch looks good to me, but i would like to have slangasek take a look before we land that (i think he looked into it before) [11:25] ljose, whats the system you run the ssh command on [11:25] Ubuntu 15.04 [11:26] (ubuntu ? ) [11:26] ogra_: :) [11:26] hmm, that should have the avahi daemon [11:26] kubuntu actually [11:26] ljose: and you got the 15.04 xz? [11:27] webdm.local is resolved using an avahi mDNS request [11:27] if both devices are on the same network and your desktop has the avahi bits installed it should just work ... [11:27] ah, kubuntu might not have that, good point [11:27] i'm not sure kubuntu installs the avahi parts ... though i wouldnt know why if they dont :) [11:28] kubuntu-desktop pulls it in [11:28] as a task [11:28] yeah, would be weird if it didnt [11:29] ljose: could you check you have avahi-daemon installed? [11:29] wait, not the daemon [11:29] hm [11:29] ljose, the first setip runs cloud-init to set up the user and ssh keys etc ... that can take a few minutes ... did you wait a while before trying it ? [11:29] libnss-mdns [11:29] ljose: ^ [11:29] Chipaca, avahi-daemon is also a client :) [11:29] Chipaca: more than 10 minutes [11:29] yeah, after 10min it should be done [11:30] /dev/sdd is the device and seems to have any partitions now after "dd" [11:31] ljose: silly question, is the bbb's ethernet plugged in to the network? [11:31] just USB [11:31] ljose: that won't work then [11:31] doesn't work like the original image over USB? [11:31] ljose: we don't bring up the usb device [11:31] ljose: no [11:31] ljose: we should, probably/maybe, but right now don't [11:32] right, by default only the onboard ethernet is brought up on boot [11:32] the usb one would be so much more convenient [11:32] let's plug the network and see [11:32] you usually plug the usb in to power the thing anyway [11:33] you might need to reboot [11:33] at least while testing :) [11:33] i'm not sure webdm will start dynamically if you now plug it in [11:33] (geez, whats up in france !! /me just watches the news on the side) [11:40] ogra_: France is a big place could you be a little more specific? [11:41] davmor2: beheading close to lyon [11:41] http://www.bbc.co.uk/news/live/world-europe-33287095 [11:55] Same problem after connect network cable, and avahi-daemon is running, any ideas? [11:55] did you reboot the BBB ? [11:55] yep [11:55] hmm [11:56] can you reach it by IP ? [11:56] ssh ubuntu@ should work too [11:57] I reboot the BBB but there isn't any blinking leds , that seems odd [11:57] just the power, and network leds [11:57] yeah, at least during boot something should blink [11:57] do you have a serial cable to check the boot ? [11:58] usb? [11:58] FTDI [11:58] something like http://www.amazon.com/Ftdi-TTL-232r-3v3-Serial-Converter-Cable/dp/B00M41OUYA/ref=sr_1_1?ie=UTF8&qid=1432041571&sr=8-1&keywords=FTDI+to+TTL+3.3 [11:59] (pretty essential when working with embedded boards) [11:59] I don't have this right now [11:59] ah, sad, would have been helpful to see whats goiing on [12:00] should I see the flash partitions when connect the USB cable? [12:00] I see /dev/sdd but no partitions there [12:01] ljose: yes, you should see four partitions [12:01] no, you shouldnt [12:01] no? [12:01] dah [12:01] we dont even enable the USB gadget [12:01] the usb cable [12:01] no [12:01] sorry [12:01] the sd card, yes [12:01] i dont get how you see /dev/sdd though [12:02] since that driver isnt on in our images it shouldnt create such a device ... [12:02] are you sure it is booting from SD at all ? [12:03] (try rebooting with the user button (S2) pressed) [12:04] I just install the image in /dev/sdd , that was the BBB device [12:04] attached to USB on my computer [12:05] uh [12:05] thats will liekyl not work [12:05] ok so I just mess the BBB image? [12:06] well, you should be able to restore the internal MMC from some debian-BBB image from beagleboard.org [12:07] i'm pretty sure you cant bot from a dd'ed snappy image from the internal MMC yet though ... needs an SD [12:08] I don't follow, I need an extra SD to use Snappy with my BBB? [12:09] yes, currently we have no way for on-board installs [12:10] What kind of SD? can I use a USB disk? [12:10] a micro SD [12:10] you could use USB but still ned the bootloader on SD i think [12:11] the initrd mounts the partitions by label ... so if you put the image on USB and have only the /boot partition on SD that would also work i guess [12:11] but you need the SD for booting [12:12] (there are surely ways to hack the /boot partition and bootloader into the MMC, but we have no automatic or userfriendly way yet) [12:13] * ogra_ will at some point take a look at rcn-ee's installed script for the BBB so we can provide an install image [12:14] so dos BBB include SD card reader? [12:14] yes [12:14] well, microSD [12:14] so I need a microSD and a reader for my PC to load the image? [12:15] to write it, yes [12:16] and if you work with embedded boards investing the $10 for a proper serial cable makes sense and saves a lot of headdache :) [12:17] ogra_: Yes I will get one, just my first BBB project [12:18] :) [12:42] meh, image build is broken [12:43] oh oh, ! damn ... i saw the mail ... thoght ... ah, arm64 again ... and ignored it [12:43] we really really need to do something about these arch names ! [12:44] +dnsmasq:*:16612:0:99999:7::: [12:44] where does that come from ? [12:44] did we change seeds ? [12:44] (doesnt look like we would want that in the core image) [12:45] * ogra_ wonders if some of seb128's changes leaked into core via seed deps or some such [12:45] ogra_: might be the addition of ubuntu-fan [12:45] ah ! [12:45] ogra_, I doubt that has to do with my changes [12:45] still. do we really want a DHCP server preinstalled ? [12:46] ogra_: if we want ubuntu-fan I don't see a way out of it [12:46] sergiusens, ogra_, btw, tried to build a personal image today for some reason I can't log in with ubuntu/ubuntu [12:46] seb128, yeah, not blaming you, rather a mis-setup of our seed deps that is exposed now ... but what sergiusens says is more likely [12:46] seb128, well, i doubt you have cloud-init installed ? [12:47] ogra_, we do [12:47] (which sets up the account on first boot) [12:47] which is creating a boot timeout [12:47] cf #ubuntu-desktop log from earlier [12:47] oh, and do you see it setting up the account ? [12:47] I guess not [12:47] it is quite wordy on console when doing it [12:48] it errors loop on url_helper warnings about trying to call http://109-04-04/...instance-id [12:48] then timeout after 90s [12:48] sergiusens, well, do we want ubuntu-fan actually *inside* core in the long term ? [12:48] so I guess that fails [12:48] i thought that was only a temporary thing [12:49] ogra_: it's like this, we either need a framework for frameworks or it needs to be in core :-/ [12:49] then we need a framework [12:50] core should really only offer an embedded os ... not fancy stuff imho [12:50] ogra_: a framework that frameworks can depend on [12:50] ogra_: and that breaks the rule for frameworks [12:50] yeah, whatever works :) [12:51] ogra_: in any case, you can revert the seed change and we can think of this [12:51] * ogra_ leaves the details to the architects ;) [12:51] heh, another problem for lool to solve [12:51] :-) [12:51] :) [12:52] I guess cloud-init is not hitting the same timeout errors on core atm? [12:52] I wonder if that's a local issue [12:52] no. it works fine [12:53] does it need to access the internet to boot? [12:53] it might expect an eth0 device [12:53] not sure what http://109-04-04 is [12:53] * ogra_ neither [12:53] but it was working on the image from a week ago [12:53] weird [12:53] i never looked at cloud-init ... i only always see the ton of messages on first boot ... [12:54] sergiusens, well, i dont think we should change seeds back right now, someone might expect fan asap ... [12:55] but we should start a discussion, imho adding stuff like dnsmasq is wrong [12:55] (what if i ship a snap using isc-dhcpd for netbooting my thin clients from my thin client server snap ... they will clash and break) [12:56] ogra_: right, I would unseed for now, no one was expecting this today [12:57] oh, ok [12:57] to not be broken over the weekend [13:03] seb128: the only reason I think your login could be broken is if for some reason the ssh host keys failed to be created and it cloud-init bailed on you [13:03] seb128: oh, and why do you have cloud-init? [13:03] because he has no user at all otherwise [13:03] sergiusens, dunno, because core has it and we started from there? [13:04] seb128, i guess you want to port the touch user creation bits for now ... and drop cloud-init [13:04] seb128: I would remove it, but we can accomodate for it, I'm not sure the 'personal' target has the nocloud seed creation going on [13:06] ogra_, sergiusens, unsure what needs changing? seeds? livecd-rootfs? [13:06] seb128, seeds to drop cloud-init ... livecd-rootfs to move over the user creation from touch [13:07] ogra_, k, let me find the user creation bits [13:07] why is core using cloud-init rather than doing it the touch way? [13:07] seb128, 01-setup_user.chroot [13:07] ogra_, danke [13:07] cloud-init brngs the ssh config [13:07] *brings [13:07] touch has ssh off by default [13:08] mvo: http://paste.ubuntu.com/11778540/ [13:09] mvo: I can push that to the ppa if you want (took me 10s to make the code change and 30 minutes to rebuild a test package :-P) [13:10] sergiusens, mvo http://paste.ubuntu.com/11778546/ ... that would fix the build error [13:11] but not sure if we want that ... or instead revert the seeds and have a proper discussion first [13:12] sergiusens, not sure you saw my ping above ... technically it looks correct to me, but i would feel better if someone like slangasek could take a look and sign it off ... he has looked at adduser before [13:12] (and might know things we dont) [13:12] ogra_: I saw it, but slangasek will only be avail in 3 hours :-) [13:13] (there are twi whitespace changes you might want to fix btw) [13:13] ogra_: and I was just fixing something mvo pointed out :-) [13:13] *two [13:13] line 135/136 and 144/145 [13:14] but thats just nitpicking :) [13:14] sure [13:14] but this quilting business is so tiresome :-P [13:14] heh [13:17] ogra_: there, vorlon has been _mentioned_ in the card [13:17] :) [13:19] ogra_: just for you http://paste.ubuntu.com/11778568/ [13:19] haha, thanks :) [13:19] so much shorter all of a sudden :P [13:20] ogra_: paste needs to learn how to color patches with a patch ;-) [13:21] i'm surprised /var/lib/extrausers/shadow is already in defines.h [13:22] ogra_: that's added in debian/patches/1010_extrausers.patch [13:23] ogra_: here you go http://paste.ubuntu.com/11778582/ [13:23] aha, mterry [13:24] ah, thats for the phone [13:24] (i think i even signed it off back then :P ) [13:27] fgimenez, elopio, balloons: https://wiki.ubuntu.com/Snappy [13:27] (and subsequent pages) [13:27] feel free to change/add as you see fit [13:27] oldschool \o/ [13:28] now we just need a gdoc -> wiki importer :) [13:29] haha [13:29] * mvo is in a meeting [13:29] poor guy [13:30] dholbach, woot [13:34] ogra_: I can make the extrausers thing without the toggle thing as well if you want [13:34] no, the toogle is good i think [13:35] sergiusens: I'm not a fan of that idea [13:35] the ubuntu-fan framework of frameworks [13:35] lool: no one is [13:35] lool: so in core it is for now [13:36] lool, i'm not a fan of shipping a dhcp server :) [13:36] (in core that is) [13:36] I'm not a fan of fan in the first place [13:37] we might want to consider cloud vs core seeds for snappy [13:37] i assume fan has its fans in the cloud :) [13:37] a fan to move the cloud [13:37] fantastic idae [13:37] lol [13:38] :D [13:51] sergiusens: "++ (void) fputs (_(" -E, --extrausers Use the extra users database\n"), usageout);" the -E is still there in the help output [13:53] sergiusens, oh, line 76 could use indentation fixing [13:53] hmm, 83 to 85 too [13:54] yay for tabs vs spaces [14:16] good morning. [14:16] thanks dholbach! [14:17] do you want to hangout today fgimenez ? [14:18] elopio, as you like, don't have too much to say, all the merges are already done [14:19] fgimenez: thanks for that. [14:19] I'll merge yours too. [14:19] elopio, ok thanks [14:19] fgimenez: lets skip today. I talk too much, you need some free time from me :) [14:20] elopio, ok np :) [14:20] ogra_: hmm, mvo left, the latest diff didn't have that -E [14:20] ogra_: and the whole file is mixed [14:29] ogra_: hmm, [14:29] hmm [14:29] I'll fix all of those [14:37] ogra_: mvo http://paste.ubuntu.com/11778888/ (the -E was already fixed in diff you missed) [14:37] * sergiusens raises fist at nothing (the lack of vcs) [14:38] sergiusens: ta! [14:38] sergiusens, heh, but the whitespace change on the two comment lines is back ... [14:38] mvo: and I was thinking if maybe this should be an implicit thing now (more invasive diff though) [14:38] just leave it though before it gets messed up again [14:38] ogra_: oh, I need to manually remove those :-P [14:40] ogra_: there you go http://paste.ubuntu.com/11778897/ [14:40] heh, thanks :) [14:40] mvo: what do you think of implicit vs explicit? [14:41] sergiusens: hm? sorry, lacking context (plus meetings) [14:42] sergiusens, as long as it is *plicit :P [14:42] mvo: instead of -E, if /etc/passwd and friends fail to lock to try and use the extrausers ones [14:43] * ogra_ wouldnt do any automation here ... in case we want to upstream it [14:43] it is more likely to be included if it is swiatchable [14:43] -a [14:43] sergiusens: explicit sounds better to me [14:44] sergiusens: do we need adduser at all (asking, no idea if yes or no) :) [14:44] well, we need some tool to add users ... [14:44] if we dont use the default (adduser) as documented everywhere for ubuntu, we need to have specific docs explaining how to add users then [14:45] personal might want it though [14:45] for multiuser desktops [14:54] mvo: ok, let's keep it explicit, and we need it for cloud init as well [14:54] ok then, we are good to go then [15:02] sergiusens: ok [15:02] sergiusens: ta [15:09] pitti, so this seed change ... is that a hard req now ? [15:10] * ogra_ feels rather uneasy about it [15:22] sergiusens, seed revert done ... (i'll wait one publisher run before kicking the image, not sure when the task generation happens) [15:35] sergiusens: mvo: elopio: fgallina: http://npf.io/2015/06/testing-exec-command/ [15:36] cool, thanks Chipaca. [15:36] I'm puzzled about this: http://paste.ubuntu.com/11779125/ [15:37] the build works with the default arch, but fails using GOARCH=arm. [15:38] elopio: you need more than just goarch, look at webdm's build script [15:38] * elopio looks. [15:48] elopio: or alternatively read doc/cross-build.md [15:48] sergiusens: ^ [15:50] Chipaca: hah, that's the second time you tell me that. Sorry. [15:50] * elopio bookmarks the cross-build. [15:50] :) [15:51] I was missing one of the packages. [16:09] nice weekend everyone o/ === seb128_ is now known as seb128 [16:52] ogra_, sergiusens: I didn't look at adduser, I only pointed in its direction. what input did you need from me on it? [16:54] slangasek, just a glance over http://paste.ubuntu.com/11778897/ if you see anything insane [16:54] technically it looks ok to me ... i'm just not sure about "conceptual" :) [16:57] sergiusens: this grub.cfg should work with both old and new and do the right thing and etc: http://pastebin.ubuntu.com/11779470/ [16:58] sergiusens: however, i haven't been able to test the upgrade [16:58] sergiusens: is there a way of testing this without publishing an image with this grub? [16:59] sergiusens: super simple the grub.cfg in the end, and i had all the bits already [17:00] sergiusens: more DRY: http://pastebin.ubuntu.com/11779488/ [17:01] sergiusens: and now i'm off to make pizza. will check back later wrt the above. [17:09] ogra_: ah, so changing shadow instead of adduser. hmm. I think it's a better logical fit in adduser, but clearly doing it in shadow would work [17:09] well, i think sergiusens plans to do both [17:10] but have shadow do all the work [17:10] so adduser needs just the switch enabled [18:09] Chipaca: nice [18:13] Chipaca: ogra_ what if we allow expansion of cmdline with a grubenv? [18:23] slangasek: my worry is that some component in some place will need to do an 'if RO' or 'if snappy' and apply the right switches; is that useradd or cloud-init? [18:24] sergiusens: I'd suggest that it should be part of the adduser config [18:25] there's not currently an option for this in /etc/adduser.conf, but it could be done [18:31] slangasek: ok, I was thinking of that and don't mind adding it there; thanks [18:32] sergiusens: ok, great :) [20:39] elopio, trying to update to latest snappy edge is giving me an error; if I use snappy update, rather than snappy update ubuntu-core: http://paste.ubuntu.com/11780433/ [20:40] as an aside, will snappy support diff based updating? Or will I always have to grab the entire image? [20:56] slangasek: can you look at https://code.launchpad.net/~sergiusens/ubuntu/wily/adduser/extrausers/+merge/263169 later? [21:12] sergiusens: ok, queued for looking [22:36] balloons: please report a bug for that, I'll try to reproduce it. [22:37] balloons: I think you are downloading only the delta. === tt_ is now known as Guest36352 [22:43] trying out snappy on the raspberry pi 2, i installed docker and booted a container with "docker run -it -p 4444:4443 nblumoe/rpi-clojure". I installed netcat and ran "echo yo | nc -l 4444". Should I not be able to "nc 127.0.0.1 4443" and get "yo" from the pi? [22:43] i can see the port bound with docker ps "a1cee09fade8 nblumoe/rpi-clojure:latest "bash" 14 minutes ago Up 14 minutes 0.0.0.0:4444->4443/tcp fervent_babbage" [22:46] but no cigar [23:42] Guest36352: format: ip:hostPort:containerPort | ip::containerPort | hostPort:containerPort | containerPort [23:43] Guest36352: so 4444 is the hostPort 4443 is the containerPort [23:43] so you need to listen to 4443 in the docker container [23:43] and connect to 4444 from the host