=== chihchun_afk is now known as chihchun
murphylaw /msg nickserv register clobrano irc*10QpAlZm06:22
dholbachgood morning06:54
fgimenezgood morning07:14
admcleod1hi guys, deplying docker containers in bridge mode, looks like im hitting a veth (or mac address? limit) seems to be ~50: systemd-udevd[5378]: Could not generate persistent MAC address for veth8aed386: No such file or directory08:33
admcleod1is it code or config?08:33
JamesTaitGood morning all; happy Cheesecake Day! 😃08:50
vmayoralogra_: ping09:16
ogra_hey vmayoral09:16
vmayoralogra_: greetings :), could you please point out which config file you guys use to compile BBB kernels?09:17
ogra_vmayoral, oh, thats ppisati  land :)09:17
vmayoralogra_: great, thanks for referring09:18
ogra_effectively it should be in /boot by default though09:18
ogra_(BeagleBoneBlack)ubuntu@localhost:~$ uname -a09:18
ogra_Linux localhost.localdomain 3.19.0-23-generic #24-Ubuntu SMP Tue Jul 7 18:56:34 UTC 2015 armv7l armv7l armv7l GNU/Linux09:18
ogra_(BeagleBoneBlack)ubuntu@localhost:~$ ls /boot/config*09:18
vmayoralppisati: true! found it, thanks a lot09:18
vmayorali was fighting with "omap2plus_defconfig" but didn't looked there. Great, thanks!09:19
ogra_make sure the version matches ... that only works on installs that havent seen a kernel upgrade yet09:19
ogra_(its a bug, we need to actually ship the config in the a/b subdirs)09:19
ogra_pitti, hey ho ....09:20
ogra_pitti, in the systemd world, which parts of the system are responsible for fstrim ?09:28
pittihey ogra_09:28
ogra_we moount with discard by default currently ... but that seems to be a no-op on most SDs09:28
ogra_(and i bet on many USB sticks too if it comes to x86)09:28
pittiogra_: it's not really init system specific -- util-linux has /etc/cron.weekly/fstrim which regularly calls it09:29
* ogra_ checks if we ship that09:29
ogra_yeah, we seem to09:29
pittiogra_: it seems regularly calling fstrim is preferred over "discard", at least on desktop-ish use cases09:29
pittiit's a "performance sucks once a week" vs. "performance sucks a little all the time" tradeoff mostly09:29
ogra_pitti, yeah, discard comes from our phone install09:29
pittibut at least in the past there were also some actual bugs with discard09:29
ogra_the MMCs fully support it so we preferred to not have a userspace process consuming extra cycles IIRC09:30
pitti(back in 20 mins)09:30
ogra_for snappy we should drop the discard though i think09:30
pittiogra_: yeah, the tradeoff might be different on a phone; discard might make sense there, then the cronjob ought to be a no-op (and we could even disable it)09:30
ogra_and rely on fstrim09:30
pittiI agree09:31
pittisnappy usually runs all the time, phones are often suspended etc, so cron jobs don't work well on a phone09:31
pitti... until we snappify the phone of course :)09:31
ogra_bug 147971109:40
ubottubug 1479711 in initramfs-tools-ubuntu-core (Ubuntu) "snappy should not mount with "discard" option" [Undecided,New] https://launchpad.net/bugs/147971109:40
ogra_pitti, ^^^09:40
longsleepyay +109:41
* ogra_ will collect some more freedback and then just drop it from the initrd if nobody objects09:42
ogra_ooooh ... new splash screen on my phone, nice !09:45
longsleepquick snapcraft questin - can i just add a plugins into my project or do i need to att them to snapcraft plugins folder?09:57
clobranohi guys, I'm trying to package some simple utilities like minicom/wget, but I don't get how to declare their shared libraries dependencies09:58
ogra_you dont, you ship them in the snap09:58
ogra_under usr/lib/<arch_triplet>/09:59
clobranoI though it was like this, but still get errors like "error while loading shared libraries"09:59
clobranoogra_: ah sorry, didn't read the last message, I'll try that10:00
=== akarasulu_ is now known as akarasulu
longsleepanyone can point me on an example snap which starts a service with a configuration file, preferrably created on first start or something?10:32
* longsleep wonders if he is the first one who needs to do that10:32
clobranoogra_: thank you, that worked10:57
Chipacalongsleep: what do you mean?11:19
Chipacalongsleep: you've got $SNAP_APP_DATA_PATH11:21
Chipacalongsleep: you can write to that, if that's what you mean11:21
Chipaca(and if you can't, that's a bug :) )11:21
longsleepChipaca: yes - i learned a lot about this the last hour - though i was wondering if i have to write all the scripts myself or if there is some generic scripting somewhere to do this11:21
longsleepi mean, i was looking for something similar to the postinst preinst stuff from debian packages11:22
Chipacalongsleep: i'm not sure what scripts you mean, here11:22
Chipacano such thing11:22
Chipacaso, atm, yes you write your own thing11:22
Chipacai'd be glad to help11:23
longsleepall right - thats what i figured - so i was wondering if someone else already did something11:23
longsleepany help would be very welcome, i can easily script the stuff but i would prefer to know something about why my script is started11:24
longsleepwas it the first install11:24
longsleepwhat was the old version11:24
longsleepstuff like that11:24
Chipacayes, i was just thinking about that a bit11:25
Chipacai don't have anything particularly insightful yet :)11:25
longsleepalso there seems to be a folder "current" inside /writable/system-data/apps/spreed-webrtc/11:25
Chipacabut i suspect a lot of people will need to have something like a "run once" or "run on version change" script11:25
longsleepcan you tell me if that is always the case?11:25
Chipacalongsleep: if the current symlink isn't there, the app isn't "installed"11:26
Chipacalongsleep: (you can have it present but not installed)11:26
longsleepChipaca: ok - so i should always use the current path in anything i need to access from there?11:26
Chipacalongsleep: anything where you don't care about the version, yes11:27
longsleepok that is helpful already :)11:27
Chipacaogra_: do you know if Bad Things happen if I reference a mounted partition from g_multi?11:27
Chipacalongsleep: note you've got a bunch of environment variables to help you with this11:27
Chipacalongsleep: hello-world.env prints them out11:27
Chipacathe ones starting with SNAPP_ are deprecated ;)11:28
longsleepChipaca: yes - i saw that one - i will check it out11:28
longsleepChipaca: oh - all right so i will try to avoid those11:28
longsleepChipaca: what about calling system utilities in general, how should i call to those, full path, expect them to be in $PATH, do not call them, ship them?11:29
longsleepfor example openssl11:29
ogra_Chipaca, hmm, for what usecase with the g_milti module ?11:29
Chipacaogra_: i'm testing it out on snappy bbb; main thing i want is to have network (& serial?) but if it also gives me safe access to a partition that's a win11:30
Chipacalongsleep: try it and tell us? it should work unless you're reading/writing things you shouldn't (or we forgot about :) )11:31
ogra_you want to expose it as usb device to be able to mount it ?11:31
Chipacaogra_: g_multi lets me do that, apparently. I'm asking if it's safe :)11:31
ogra_that should be possible ... the MMC debian install does it, i bet just stealing the config from there would work ;)11:31
longsleepChipaca: so the question is what i can expect for tools like sed, grep, awk, openssl and possibly other stuff. I mean for debian packages i have dependencies on those. What if one of these tools are suddenly gone in newer images?11:32
ogra_(since that seems to be tested and proven already)11:32
Chipacaogra_: but the debian mmc seems to reference the sd card11:32
* Chipaca looks11:32
ogra_and you dont want that ?11:32
Chipacasudo modprobe g_multi bcdDevice=0 cdrom=N file=/dev/mmcblk0p1 host_addr=D0:5F:B8:A3:92:81 iManufacturer=Circuitco iProduct=BeagleBoneBlack iSerialNumber=C0-3614BBBK6053 idProduct=0 idVendor=0 luns=0 nofua=Y num_buffers=2 qmult=5 removable=Y stall=N11:32
ogra_i think it even only references /boot currently11:32
Chipaca^ that's what debian does, translated into modprobe11:33
Chipacai need to boot back into debian to see if mmcblk0 is the sd card or not11:33
ogra_yeah, so just copy it and use whatever file= option you like :)11:33
Chipacalongsleep: sed, grep, awk should work11:33
ogra_i think it is the MMC in debian11:33
Chipacalongsleep: openssl should work afaik11:33
Chipacalongsleep: when i say "should" i mean "it's a bug if it doesn't", and also that we're using some of those in e.g. hello-world stuff11:34
longsleepChipaca: yes they all work at the moment, though i could not find a list of "save to use tools from the system". It is probably something which should be created.11:34
Chipacalongsleep: that is: you can execute anything in the usual system path11:34
Chipacalongsleep: some things, like python :), will go away at some point (from core)11:35
longsleepChipaca: right, but how do snaps handle it when they depend on a tool which went away?11:35
Chipacalongsleep: we might at some point go crazy and replace all the rest with busybox, but it's unlikely :)11:35
Chipacalongsleep: that's a good question11:36
Chipacalongsleep: the quick and wrong answer is that they shouldn't depend on things that're going to away11:37
longsleepright :)11:37
Chipacathe true answer is probably "we need to write some kind of thing that promises what won't go away"11:37
Chipacalongsleep: for a start, everything POSIX asks for, will be there always11:38
Chipacalongsleep: that leaves out, i think, awk, perl, and python11:38
Chipacaand iptables, say :)11:39
longsleepChipaca: Perl and Python should be frameworks if you ask me11:39
Chipacalongsleep: well, except frameworks don't work like that, but yes11:39
longsleepChipaca: awk is nice to have, for my use case, i need to make sure i got openssl11:40
Chipacalongsleep: the expectation right now is that if you're building using perl or python, you should ship those in your snap11:40
longsleepChipaca: all right, thanks - lets see what i can do with some script fu11:42
Chipacaogra_: rsalveti: we need to make a list of things that are not going away :)11:42
longsleepChipaca: what about the shell, is it bash, dash, whatever else?11:42
Chipacaogra_: rsalveti: and pretty pleas let's have openssl on that11:42
Chipacalongsleep: /bin/sh is dash11:42
longsleepChipaca: ok, any plans to change that?11:43
longsleepor rather, dropping /bin/bash ?11:43
Chipacalongsleep: i don't know about bash11:44
Chipacalongsleep: in my mind, it's huge and bloated and i love it11:44
Chipacalongsleep: so once things settle and we start paring down, it's a candidate to go11:45
Chipacalongsleep: in my mind11:45
Chipacalongsleep: but note we have not discussed this at all11:45
longsleepwell, i might be able to use dash only - though i do not really like it11:45
* Chipaca is aware11:45
* ogra_ would like to see bash moved to comfy 11:46
ogra_longsleep, https://wiki.ubuntu.com/DashAsBinSh has some hints for scripters :)11:47
Chipacawe should go make our own distro, call it uubuntu (because u- is for µ; also because the long first vowel is distinctive of my Córdoba roots)11:47
longsleepogra_: yeah - i had some fun in the past porting scripts to upstart11:47
Chipacadash and busybox's ash are essentially the same, yes?11:48
longsleepChipaca: so the hello world example showing the variables you mean http://bazaar.launchpad.net/~snappy-dev/snappy-hub/snappy-examples/files/head:/hello-world/meta/ right?11:48
Chipacawith all the bling turned on in busybox that is11:49
Chipacalongsleep: i mean if you install it and run hello-world.env it'll print them out for you11:49
* longsleep has no idea about the differences between dash and busybox11:49
longsleepChipaca: right - let me try that11:49
Chipacawe could sort the output, that would be nice of us11:49
Chipacaogra_: any idea whether i can read the bbb's serial number from anywhere?11:50
Chipacaogra_: also, any idea for generating a stable random mac on the bbb :)11:51
longsleepChipaca: well anyone can |sort11:52
Chipacalongsleep: yes. But if the script did it itself, you'd know more about what tools you can use :)11:52
longsleepright, what is the history of the SNAPP double P vars?11:53
ogra_Chipaca, no, i just noticed it isnt set ... usually from /proc/cpuinfo11:53
Chipacalongsleep: snappy app -> snapp11:54
Chipacalongsleep: somebody thought it was a good idea11:54
ogra_with the latest image you should definitely have afixed MAC though11:54
Chipacalongsleep: they changed their mind :)11:54
ogra_i surely do here11:54
Chipacaogra_: i meant for g_multi11:54
Chipacaogra_: because the bbb has *two* ethernets! one is hidden inside g_multi :D11:55
* Chipaca loves all the ethernets11:55
ogra_hmm, probably there is something hidden in sysfs11:55
Chipacaogra_: i'll grep11:55
longsleepfor the ODROID, the serial is available in U-Boot and could be passed as parameter to the kernel.11:56
longsleepmaybe it is similar with the bbb11:56
Chipacalooks like it11:57
ogra_yes, i just dont get why it doesnt end up where it should11:57
* ogra_ will have to dig11:58
* Chipaca just mounted /writable on his host11:58
davmor2ogra_: I don't know how often I can explain this, but technology hates us, you have to hate it back, it doesn't help the issue but it makes you feel a lot better ;)11:59
Chipacaogra_: i could also just let g_multi assign random things, that works11:59
ogra_well, i dont really think it matters since the USB cable will never see a real ethernet :)12:00
ogra_so theoretically your MAC can be whatever you want it :)12:00
Chipacaogra_: it only matters if you're doing fancy firewally stuff12:00
ogra_doews that MAC matter for firewalling over USB ? i'd expect that to hook into a higher layer12:01
ogra_it might matter for IP assignment12:02
longsleepChipaca: so where should i create the config file, in the environment i only see folders which contain the version. I need to store stuff which is independent from version.12:03
longsleepChipaca: i see this SNAP_APP_DATA_PATH=/var/lib/apps/hello-world.canonical/1.0.1812:03
longsleepbut i would need /var/lib/apps/hello-world.canonical/12:03
longsleep(assuming it is writable)12:04
Chipacalongsleep: the data path is always versioned12:04
longsleepChipaca: mhm so where do i store stuff i want to keep across versions then?12:04
Chipacalongsleep: you should have /var/lib/apps/, and it should be writable by your service12:04
Chipacalongsleep: on upgrade, the data directory is copied12:04
=== chihchun is now known as chihchun_afk
longsleepChipaca: ah ok12:05
Chipacalongsleep: service stopped, data directory copied, current symlink redirected, service started12:05
longsleepChipaca: sounds good thanks12:05
Chipacalongsleep: um12:05
Chipacalongsleep: if the user rolls back, and then rolls forward, things get iffy12:06
=== chihchun_afk is now known as chihchun
ogra_seems uboot has it12:06
ogra_how do we get it across to the kernel now ?12:06
longsleepjust add some parameter and parse /proc/cmdline in user space12:06
longsleep(that is how android does it)12:07
ogra_well, i would be hoping there is no need to parse anything if we use the right cmdline option12:07
longsleepogra_: yeah - you might have a driver in place which does it already12:07
Chipacaogra_: sudo modprobe g_multi bcdDevice=0 cdrom=N file=/dev/mmcblk0p4 iManufacturer=Circuitco iProduct=BeagleBoneBlack idProduct=0 idVendor=0 luns=0 nofua=Y qmult=5 removable=Y stall=N12:08
Chipacaogra_: check if that gives you the right mac (in dmesg)12:08
Chipacai wonder what it used iSerialNumber for; it isn't complaining without it12:09
Chipacaogra_: i say we add that to bbb's modprobe.d :)12:09
* ogra_ twiddles thumbs waiting for cloud-init again 12:11
Chipacathen i can ditch the clumsy ethernet cable! woo :)12:11
Chipacagrmbl grmbl cloud-init12:11
ogra_hmm, no, the MAC doesnt match anything from uboot12:12
sergiusensChipaca: implement https://trello.com/c/W5WiZQM7/117-core-config-for-cloud-init ;-)12:12
Chipacaogra_: ah well12:13
Chipacaogra_: i still say we add that to the modprobe12:13
Chipacacan figure out the serial number and mac later12:13
Chipacait doesn't seem like anything breaks without 'em12:13
Chipacanote that modprobe lets you mount writable on the host12:14
Chipacawhich is pretty sweet, if you ask me :)12:14
Chipacasergiusens: i'd love to12:14
Chipacarsalveti: ^ put me down for core-config-for-cloud-init when i return from my holidays12:15
Chipacapretty please :D12:15
Chipacathings i learned while having lunch and looking into mac addresses: osteodiastasis is nasty12:28
rsalvetiChipaca: sure :-)12:31
elopiogood morning.12:33
rsalvetielopio: good morning12:37
=== chihchun is now known as chihchun_afk
longsleepAnother snapcraft question, any ideas if someone tried to do cross snapcraft, eg. create a snap for armhf on amd64?13:12
* longsleep wonders how multi arch snaps work anyway13:13
ogra_longsleep, i guess thats a mterry or ted question13:14
ogra_(like all snapcraft questions somehow :) )13:14
ogra_both are in US TZ ...13:14
mterrylongsleep, I'm up13:15
longsleepgreat :D13:15
tedlongsleep, We're gonna initially target people building an arm snap on arm.13:15
tedWhich isn't ideal, but that's phase 1.13:15
mterrylongsleep, plan right now is you build on the target platform.  Future plan is to use comfy to make that much easier13:15
longsleepted: all right, i guess i can do some chroot stuff then13:15
ogra_the magical comfy13:15
longsleepmterry: i still do not know what comfy is13:15
* ogra_ adjusts his macros from "snappy will fix everything" to "comfy will fix everythng" :)13:16
mterryogra_, :)13:16
longsleepted, mterry, ogra_ and what about the multi arch. Do i have to build multiple snaps for each arch and somehow can upload them all into the store?13:16
mterrylongsleep, it's basically "normal Ubuntu embedded in Snappy for developers' benefit"13:16
ogra_longsleep, i think thats possible, yup13:17
longsleepok - so the file name of the snap should include the architecture?13:17
mterrylongsleep, the current only way to do that is embed multiple binaries in one snap (which is gross and tooling isn't great).  Future plan is to allow uploading multiple snaps, one per arch, for one store item.  Or you could upload them as separate store items now13:17
ogra_(i personally prefer one snap for all arches, but then you cant use snapcraft atm as i understand)13:17
longsleepoh - so i have to put them all into one snap :/ that is unfortunate13:18
ogra_no, you dont13:18
mterrylongsleep, it's a store limitation right now13:18
sergiusensogra_: you flip flopped the preference it seems ;)13:18
mterryogra_, did store grow support for multiple snaps?13:18
ogra_sergiusens, after using the model, yes :)13:18
sergiusensogra_: it is easier in some aspects, just not ideal for huge snaps13:19
ogra_mterry, hmm, not sure about package namin, but you can surely upload for a specific arch ... would likely have to be $snapname-$arch.snp then though13:19
ogra_i know i can upload armhf only or amd64 only for one snap13:20
ogra_sergiusens, indeed, size matters :)13:20
mterryogra_, yeah but you have to upload each arch as a separate store item right now13:20
mterryogra_, which is gross is all  :(13:20
longsleepwhat does it mean separate store item? Can it use the same name then?13:20
ogra_yeah, surely not idea13:20
mterrylongsleep, I'm meaning separate names when I mean separate store item13:21
mterrylongsleep, store can't handle multiple arch-specific snaps for same name yet13:21
ogra_longsleep, myshinysnap-armhf_1.2.3.snap ... and myshinysnap-amd64_1.2.3.snap13:21
longsleepyes file names are good, though regarding the store i am unsure13:22
ogra_that will be two separate store items with separate metadata pages in the store13:22
longsleepis there an example in the store already?13:22
ogra_(and two different places to upload to etc)13:22
longsleepi think i can be ok with that13:22
longsleepbut probably will only create armhf then for now13:23
longsleepmterry: how do you folks manage that for webdm? Does it have a different name for other archs?13:26
mterrylongsleep, I don't know...  I don't manage that myself13:27
=== chihchun_afk is now known as chihchun
mterryrsalveti, tried old BBB image, still didn't work!  So it's either user error, messed up SD card, or messed up BBB (though all are new...).  I'm going with user error...  Trying my raspberry pi2 just in case13:34
sergiusensmterry: after dd'ing make sure records out and in match ;-)13:38
mterrysergiusens, interesting...  but looking at scrollback, it seems fine13:40
Chipacalongsleep: i like your case, btw :)13:58
rsalvetimterry: makes no sense14:00
rsalvetiguess we can only know for real with a serial console14:01
rsalvetilike http://elopio.net/en/node/104214:01
rsalvetimterry: is yours also bbb rev c?14:02
rsalvetishould have a bit C in one of the stickers14:02
ogra_mterry, can we see your serial output during boot ?14:02
Chipacasergiusens: where was it golang-uboot-go-dev was at?14:12
Chipacaand how do i add that to a schroot14:12
sergiusensChipaca: in the ppa:snappy-dev/image14:14
sergiusensChipaca: https://wiki.ubuntu.com/SimpleSbuild#Temporarily_adding_PPAs14:15
Chipacayep, just found that14:15
sergiusensChipaca: you could add it permanently into a snappy branded schroot as it is the base of our builds anyways14:15
Chipacayep, am doing so :)14:16
ogra_longsleep, https://insights.ubuntu.com/2015/07/30/spreedbox-most-private-video-chat-and-file-exchange/14:20
mterryrsalveti, ogra_: sorry was in meeting about this MIT demo.  rsalveti: it is rev C.  ogra_: I've moved on to trying with the raspberry pi2 but can go back and try to get more output later14:31
rsalvetiogra_: nice14:31
rsalvetisame rev we all use, weird14:32
rsalvetiyeah, but you'd need a serial cable around14:32
ogra_well, we should somehow make clear that a serial cable is kind of expected on embedded HW :)14:33
* ogra_ gives up trying to re-share rsalveti's Win10 post on the phone and does it from the PC 14:34
ogra_unbelivabe ... i cant switch back and forth between browser and G+ to actually type one sentence off the website into my post14:35
longsleepogra_: very nice thanks!14:35
ogra_either G+ or the browser goes white14:35
* ogra_ wants 4GB on phones :P14:35
ogra_(that will at least allow me two apps at the samer time :P )14:36
Chipacaogra_: unless one of them is a browser ;)14:36
ogra_yeah, then you need 8GB14:37
longsleepis there any phone which actually can use more than 3GB ?14:37
longsleepi read somewhere that the oneplus 2 should have 4GB14:37
sergiusensogra_: yeah, I have this problem when attaching a photo through the content hub as well, the app that requested the content hub ceases to exist on return :P14:37
ogra_i dont think there are actual 64bit phones ...14:37
ogra_the 64bit ones i know all run in a 32bit compatibility mode14:38
ogra_which would limit your to 3G14:38
sergiusensogra_: mem management is atrocious on the phone lately14:38
sergiusensogra_: I don't use g+ on it anymore, too depressing14:38
ogra_sergiusens, well, i see people complaining about that photo thing all the time ... works for me ... but then i only use G+ to share stuff14:38
ogra_(and very rarely telegram)14:38
sergiusensogra_: right, my use case is untappd (take photo and return to nothing)14:39
ogra_(and indeed i use my G+ app, not that official canonical thing :) )14:39
sergiusensogra_: now I take the photo just in case I lose it ;-)14:39
sergiusensand then open the hub14:39
ogra_ah, so you share from the camera app14:39
ogra_i always pick gallery (and make sure the camera is closed when i do)14:40
ogra_seems thats less memory hungry14:40
sergiusensogra_: oh, good point14:40
mterryogra_, rsalveti: now...  with pi2, my first ssh attempt is reset by peer and the second hangs.  :-/  but that's further than bbb...14:50
ogra_mterry, and you dont have a serail cable ?14:55
mterryogra_, no, I'm connecting over local network (ethernet on pi2 and wifi on laptop)14:55
ogra_ah, well, that wont give us many meaningful infos ...14:56
ogra_try pulling the syslog off the writable partition (but i guess thats to late to catch boot inssies)14:56
mterryogra_, but might help for pi2... i could get openssh logs14:57
ogra_well, you can just pull the log from the SD directly14:58
mterryogra_, that's what I meant, get the openssh logs from the SD14:58
mterrysince I can't ssh in14:58
ogra_well, syslog would be more interesting ... the only "ssh log" we have is auth.log14:58
ogra_(and that usually hasnt mugh meaningful to debug anything beyond login issues)14:59
=== erkules_ is now known as erkules
mterryogra_, http://paste.ubuntu.com/11967112/15:00
ogra_mterry,  looks like cloud-init didnt generate the keys properly15:00
mterryogra_, is something like a "USB to Serial adapter" what I want?15:25
ogra_one sec15:25
ogra_that is what i got here15:25
ogra_you want one with the flying cablöes at one end so you can use it on boards that have different order15:26
ogra_(instead of a fixed 4-prong thinie)15:26
mterryogra_, haha, little tentacle cables.  OK15:26
ogra_yeah :D15:27
mterryogra_, will search for one at my local store15:27
fgimenezChipaca, can you give me a hand with a failing service test?15:39
fgimenezChipaca, not sure if fails the service or the test :)15:39
fgimenezChipaca, here's the error
fgimenezChipaca, you can search for FAIL:15:40
fgimenezand here's the test's code http://bazaar.launchpad.net/~snappy-dev/snappy/snappy/view/head:/_integration-tests/tests/installFramework_test.go15:40
fgimenezChipaca, it fails eventually, usually all is fine15:42
Chipacafgimenez: could you pastebin console? i don't have my vpn setup working for some reason15:44
fgimenezChipaca, sure http://paste.ubuntu.com/11967413/15:45
Chipacafgimenez: can you get more logs than that? it looks like your app is panic'ing15:46
fgimenezChipaca, yes, that's because this failed assert http://bazaar.launchpad.net/~snappy-dev/snappy/snappy/view/head:/_integration-tests/tests/installFramework_test.go#L8615:47
fgimenezChipaca, the problem is the output of "systemctl status docker_docker-daemon_1.6.1.002.service"15:47
Chipacafgimenez: is this with snappy trunk?15:48
fgimenezit expects to find the service active15:48
fgimenezChipaca, yes15:48
fgimenezChipaca, the test usually passes but 10% of the time fails15:50
Chipacafgimenez: is network up on the kvm?15:50
Chipacafgimenez: is this recent?15:50
Chipacafgimenez: wrt "network up", i'm asking if there is a default route, fwiw15:50
fgimenezChipaca, the network is up, it is being tested against a 15.04 image, and the code has been there for some time15:51
Chipacafgimenez: i mean, are the failures recent15:51
Chipacabut if it's 15.04 then it's not what i thought it might be15:51
fgimenezChipaca, not afaic, we are checking if the service is up after a reboot, maybe we are asking for it too soon?15:52
Chipacafgimenez: give me a sec15:53
Chipacasystemctl show -p ActiveState network-online.target15:53
Chipacafgimenez: delay the check until that^15:53
Chipacafgimenez: returns "ActiveState=active"15:53
Chipacafgimenez: and let me know how it goes15:53
Chipaca(even better if you wait a bit *after* it got to "active")15:54
Chipacafgimenez: if you need to cap it, that should not be more than say 3 minutes after boot15:54
fgimenezChipaca, ah! ok that makes sense, i'll try that thanks a lot! :)15:54
Chipacaas even in the worst case the network-online thing will only wait 2 minutes, as things stand15:54
Chipacafgimenez: also15:55
Chipacafgimenez: this check might be overkill right now (maybe with a sleep you'll have enough)15:55
fgimenezChipaca, yes, i was going to ask you about that15:55
Chipacafgimenez: but there'll be code in 15.04 soon that will break if you don't wait for this, so you might as well do it now15:55
Chipacafgimenez: but, your choice, really15:55
fgimenezChipaca, ok i'll go for checking the state, thanks!15:57
=== chihchun is now known as chihchun_afk
ogra_rsalveti, hmmm ... i was just looking closer at mterry's log ....16:43
ogra_Feb 16 20:55:26 localhost systemd[1]: Starting Initial cloud-init job (metadata service crawler)...16:43
ogra_Jul 30 14:41:40 localhost systemd[1]: Time has been changed16:43
ogra_Jul 30 14:41:42 localhost cloud-init[722]: Cloud-init v. 0.7.7 running 'init' at Mon, 16 Feb 2015 20:55:35 +0000. Up 51.59 seconds.16:43
ogra_seems the time gets updated in the middle of cloud-init runnign16:43
rsalvetithat is indeed interesting16:44
ogra_and cloud-init still keeps the old time cached somehow16:44
ogra_it is also weird that the last mount time doesnt seem to have been updated between the two boots ... they both start at feb 16th16:45
ogra_rsalveti, hmm, did you do a vivid build tonight ?17:02
* ogra_ just notices an arm64 build failure mail from 7am this morning for a vivid build17:03
ogra_i wonder if someone played with the crontab commants on nusakan17:03
admcleod1so.... any ideas why i cant create more than 50 veth's with docker in snappy on an rpi2? :}17:32
ogra_kirkland, any idea ? ^^17:33
ogra_(or any idea whom to point admcleod1 to ? )17:33
longsleepis there a way to directly use git with launchpad?17:39
ogra_longsleep, yeah, i think so ... ask in #launchpad though17:41
* ogra_ still uses bzr if he can choose :) 17:41
longsleepright, i want to add a simplified plugin to snapcraft which can directly use a .deb file17:42
mterryogra_, rsalveti: ok, back from a long trek to find serial cables and lunch.  Everywhere is out of them, so I ordered an overnight amazon cable17:59
* mterry will try again on current devices / work on other aspects of presentation18:01
ogra_mterry, yeah, lets look tomorrow then ... but see above, there is something going on with the clock and last mount time it seems18:02
mterryogra_, huh, I don't know what to do with that information...  :)  but interesting clue18:03
mterryogra_, does ethernet need to be plugged in at boot for it to work?18:03
ogra_mterry, nothing ... lets inspect it deeper tomorrow :)18:03
mterrymaybe I can plug it in later to avoid updating time in middle of boot18:04
ogra_well, therenet is needed by cloudinit i think18:04
ogra_the thing is, cloud-init generates your ssh keys ... i'm not sure what actually happens if the clock gets changed in the middle of generating them18:05
ogra_(i'm also not sure why your last mount time (which is what fixrtc sets the system time to if there is no rtc) is feb 16th ... that kind of predates the released image)18:06
mterryogra_, I hate hardware.  Why can't everything just be ethereal bits?18:10
tedMan, I looked and look at what I did wrong. Turns out just needed to fix the expected result of the test, it changed.18:35
mterryogra_, rsalveti: my third rewritten sd card worked for pi2....18:36
mterryI don't get it, but I'm not asking questions18:36
rsalvetimterry: weeeird18:44
rsalvetiogra_: I enabled the cronjob for 15.04/edge again18:44
rsalvetisince the release is out18:44
mterryrsalveti, not confidence inspiring for sure18:44
rsalvetido you have any external usb-sd card reader?18:45
sergiusensmterry: maybe get a new sdcard?18:45
rsalvetior that18:45
mterryogra_, people have used this webcam-demo.canonical app and it works and all that?  I'm getting "GD Error: gd-jpeg: JPEG library reports unrecoverable error: Not a JPEG file: starts with 0x26 0x81" -- trying to figure out if it is the snap or something else19:14
ogra_rsalveti, what for ? the archive wont change, dailies dont really make sense for 15.0419:45
rsalvetiogra_: does change, -updates and -security19:46
ogra_well, true indeed19:46
rsalvetiI think federico tested the webcam demo with latest release, maybe elopio gave that a try as well19:47
elopioI haven't tried that one for a month, at least.19:48
* ogra_ knows that asac tested it too not to far ago19:49
ogra_but probably before QA did19:49
mterryOK...  what's the best way to run a native go arm executable?  We have golang-go for armhf.  But running under qemu dies (known bug).  Is there anything like comfy that I can fake to get an armhf deb-based system running on my board?  using docker and installing a cloud image in that?21:01
rsalvetiyeah, docker and an ubuntu image21:01
rsalvetited probably got something21:01
rsalvetisince he's working at the comfy idea21:02
mterryrsalveti, I'm having a hard time finding an armhf image that isn't our "cloud" image.  Is that fine?  ubuntu-server and ubuntu-desktop don't seem to publish normal everyday armhf images21:02
rsalvetiit should be fine21:02
* mterry has never used cloud before, but assumes it's close to server21:02
tedYeah, I think there's another image that people seem to be using, but not an official one.21:03
tedLet me find it.21:03
* ted is curious why we don't seem to have an official one.21:03
mterryrsalveti, I realized that our current Go plugin in snapcraft is amd64-only and expanding it to armhf is easy, but hard to actually test.  :)  Most of our fun examples for snappy involve Go21:04
rsalvetiyeah :-)21:04
rsalvetiand why ted is fixing that issue for us :-)21:04
mterryThree cheers for ted!  :)21:04
=== chihchunl is now known as chihchun
tedmterry, https://registry.hub.docker.com/u/armbuild/ubuntu/21:06
mterryted, I don't like downloading things from non-ubuntu.com sites  :)  feels dirty21:07
tedmterry, It is, but make sure to not put important stuff on that device :-)21:09
* mterry scp's private gpg keys21:09
* mterry needs to probably update my gpg key to something new.... I haven't done that in 15 years21:09
mterryI'm probably using md5 as the core key mechanism21:10
tedYeah, I updated mine a while ago for that reason. I choose a really big number for the new key.21:15
tedI'm sure that makes it perfect for another 20 years.21:15
mwhudsonhey can i trick someone into looking at https://bugs.launchpad.net/snappy/+bug/147873621:28
ubottuLaunchpad bug 1478736 in Snappy "tests fail with 1.5 tip" [Undecided,New]21:28
rsalvetielopio: Chipaca: ^21:44
elopiomwhudson: that's definitely not happening on trusty, because that's what our tarmac runs.21:46
mwhudsonelopio: no it's definitely not21:46
elopioand it's not happening for me on vivid, probably not either on wily because that's what federico uses.21:46
mwhudsonelopio: context https://wiki.ubuntu.com/MichaelHudsonDoyle/Go15Preparation#preview21:46
elopiomwhudson: please give us as many information about your environment as possible. I'll try to reproduce it when I return.21:46
* mwhudson edits bug summary21:46
elopiooh, I see.21:46
elopiomwhudson: I need a couple of hours in the scary outside world. I'll reply to the bug when I come back.21:47

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