[01:30] Hi, guys :) [01:32] Any news on ubuntu snappy support on the RPI 3? [02:25] jdstrand: hey === chihchun_afk is now known as chihchun [02:26] jdstrand: could you have a quick look at https://github.com/ubuntu-core/snappy/pull/635 [02:26] jdstrand: the only new thing there since we last looked at that is https://github.com/ubuntu-core/snappy/pull/635/commits/939d1d5824c8ced1b7e994cf4f2f88ad289693ba [02:27] jdstrand: I would *really* like to land this branch as it's been in a linbo for a week and I have some code that needs it [02:27] jdstrand: if I lost track of the desired file name and other templates then please correct me but I think it is okay [02:28] davidcalle: hey, sorry, lost IRC connection [02:29] davidcalle: can you re-state your request please? === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [08:26] good morning [08:39] Good morning [09:05] jdstrand: Hi are you available? [09:19] good morning [09:48] mvo, you didnt approve the kernels from the last build/upload ? === chihchun is now known as chihchun_afk [10:00] Did you heard about the bomming in belgium dammed === chihchun_afk is now known as chihchun [10:08] ogra_: hey, good morning [10:08] ogra_: I see nothing in the unapproved queue [10:19] good morning ;) [10:19] Any news on ubuntu snappy support on the RPI 3? === chihchun is now known as chihchun_afk [10:20] I heard it will be working on RPI 3 netphreak [10:20] But i'm not a canonical developer [10:21] mvo, https://myapps.developer.ubuntu.com/dev/click-apps/4739/rev/2/ https://myapps.developer.ubuntu.com/dev/click-apps/4316/rev/3/ https://myapps.developer.ubuntu.com/dev/click-apps/4284/rev/8/ https://myapps.developer.ubuntu.com/dev/click-apps/4283/rev/16/ https://myapps.developer.ubuntu.com/dev/click-apps/4283/rev/17/ [10:22] morning zyga [10:25] ogra_: aha, please click on "request for manual review", then I can see them in the review queue, otherwise they are invisible for me [10:26] ogra_: I can do that now myself [10:26] ogra_: just for the next time, that click makes it easier for me to find them [10:26] ok [10:27] i wonder how to do that for the automated uploads then :P [10:29] jdstrand: hey [10:29] jdstrand: I didn't know you were around [10:29] ogra_: I think we need to talk to jdstrand and see if we can drop the -all-root requirement [10:29] ogra_: for the squashfs, I think that is causing the issues right now because the repack of the snap can not be done because of different parameters and this causes this manual review thing [10:29] mvo, well, at least special case os snaps [10:29] jdstrand: I don't know if you saw the discussion on telegram; we were wondering about the apparmor cache, specifically what maintains the cache enabled by https://github.com/ubuntu-core/snappy/pull/635/files#diff-48ca5117699c941398b92fcd61caaed5R43 [10:30] ogra_: yes, or that [10:30] i think the --all-root was wanted (there was a bug) [10:30] Bug #1546821 [10:30] bug 1546821 in Snappy "please add -all-root and -no-xattrs to mksquashfs options" [Undecided,New] https://launchpad.net/bugs/1546821 [10:31] ogra_, yes; the bug sequence was, make snapcraft support the os snap without all-root and remove the snappy build functionality [10:31] wednesday's release will have that [10:32] sergiusens, ah, good to know [10:37] sergiusens, please leave me a few days with the removal from snappy though (even though i expect no issues with switching livecd-rootfs, if there are any we are screwed with the dailies) [10:38] ogra_, I am not removing snappy build, that is on mvo and his merry friends :-) [10:38] ah, ok [10:41] can ubuntu-device-flash generate separate .img files for each partition instead of a big img file? [10:42] ysionneau, nope, but kyrofa has a script to split it into two img files [10:42] (one for the boot stuff, one for writable [10:42] ) === dpm_ is now known as dpm [10:43] I guess fdisk + losetup + dd should do the job? [10:43] sergiusens: we still need that bit for various tests (and its ~5 lines) happy to hide it (and make it only available in the tests) [10:43] but I'd like to have a look at the script anyway :) [10:44] davidcalle, want to have a chat about snappy/snapcraft docs in a bit? or shall we talk later? [10:45] dholbach: can you do a quick one in ~20min? [10:45] yes, a quick one [10:46] mvo, well I'm not the one pushing for it https://bugs.launchpad.net/snapcraft/+bug/1546821 look at comment #8 [10:46] Launchpad bug 1546821 in Snappy "please add -all-root and -no-xattrs to mksquashfs options" [Undecided,New] [10:47] mvo, you can hide it; but personally I think our QA would be a lot better if you use snapcraft's snap instead [10:47] mvo, if it is just 5 lines it could also just be moved to the test runner code [10:48] sergiusens: fair enough, yes, I think a separate package would be good [10:48] sergiusens: anyway, let me look at this [10:48] mvo, but of most importance; people like you and me need to stop using `snappy build` [10:53] sergiusens: thats fine, can we make "snapcraft snap" not look for snapcraft.yaml maybe? this would allow us to keep using our tests [10:54] mvo, have you used snapcraft snap? ;-) [10:54] sergiusens: http://paste.ubuntu.com/15471199/ [10:55] mvo, `snapcraft snap DIR` :-) [10:56] sergiusens: thanks! http://paste.ubuntu.com/15471206/ :P I guess it wants to tell me that the architectures field is mandatory [10:56] I spy new update coming down the pipe! [10:56] Updating canonical-pi2-linux (4.4.0-1004-raspi2+20160321.17-52) [10:56] \o/ [10:56] sergiusens: I will prepare a branch that replaces snappy build with snapcraft snap [10:57] sergiusens: will snapcraft snap also run the click-reviewers-tools automatically? [10:58] Hello guys, I didn't RTFM, but I was just wondering. Can we have several snaps installed and communicte with one another? Redis snap, Libreoffice snap, all usable from a webapp snap? [11:01] davidcalle, let me know when you have time [11:02] is this expected when i update http://paste.ubuntu.com/15471238/ - to 4.4.0-1004-raspi2+20160321.17-52 ? [11:05] dholbach: I do now :) [11:06] yes! [11:20] Good morning [11:22] ysionneau, those scripts are currently specific to the rpi2 image (they need to be generalized), but they're here: https://github.com/owncloud/pi-image/tree/master/image-creation-tools [11:36] mvo, I need to make it run it; but I'm waiting for stabilization of the whole infra for that (wrt click review) [11:36] as it might just end up confusing with all the churn [11:38] * ogra_ tickles pitti with bug 1547033 ... would be nice to have that fixed before final release ;) [11:38] bug 1547033 in systemd (Ubuntu) "please allow /etc/mtab to be a pre-existing link on readonly filesystems" [Medium,New] https://launchpad.net/bugs/1547033 [11:41] sergiusens, https://github.ubuntu.com/96boards/linux ?? [11:41] (from your blogpost) [11:42] sergiusens, i assume there sneaked some wishful thinking in to take over github with ubuntu ? [11:44] ogra_, lolz [11:44] ogra_, let me fix that :-P === jospoortvliet_ is now known as jospoortvliet === chihchun_afk is now known as chihchun [12:00] sergiusens, elopio whew, so those examples were failing for everyone? [12:02] kyrofa, yes, I dug into it last night; it is related to the snappy changes that landed yesterday [12:02] sergiusens, ah, okay [12:03] kyrofa, first -> binary_name = appname if appname == snapname else "snapname.appname" [12:03] second, PWD for apps changed again [12:03] to be PWD [12:03] kyrofa, now the testing infra is broken ;-) [12:03] sergiusens, yeah, and a number of environment variables were removed, too === chihchun is now known as chihchun_afk [12:08] sergiusens, I see an opencv failure as well... trying to track that down [12:09] mvo, do you think we can do anything about the umount error messages on shutdown (i see a lot red on reboot) [12:09] ogra_: what is that? do you have some more details? [12:09] mvo, /writable and some bind mounts cant be unmounted [12:10] ogra_: hm, I check [12:10] i think we should just skip unmounting altogether, call sync and remouont ro [12:10] and not spit out an error :) [12:14] ogra_: followed up to the bug, mind having a look? [12:15] pitti, i'll adjust that and we'll see :) === chihchun_afk is now known as chihchun [12:15] ogra_: cheers [12:16] kyrofa: thx ! [12:17] sergiusens, does the opencv example work for you on xenial? I'm getting libmvec errors [12:17] ysionneau, no problem [12:18] kyrofa, well the examples tests runs them, so I hope they do [12:19] sergiusens, yeah mine failed... and it fails locally as well. I'll rerun them just to be safe [12:21] What interfaces are there available for network? [12:21] jdstrand, haha, using indent in the launcher now? [12:21] kyrofa, don't run retest this as it will fail :-) [12:21] kyrofa: yeah, I am spending some time in there :) [12:22] sergiusens, :( [12:22] mvo, ogra_: where do you want to drop -all-root? [12:22] jdstrand, for os snap builds [12:23] (we need the system-user-owned files and dirs there) [12:24] background: -all-root is required to verify snaps from untrusted sources [12:24] otherwise I can't resquash [12:24] an os snap is of course trusted [12:25] I'm pretty sure if I remove -all-root from the resquash we are going to get different users at some point, if not today [12:25] but I can try [12:25] ogra_: what about the kernel snap? [12:26] jdstrand, there it doesnt matter .. nor for the gadget [12:26] sergiusens, ah, I was able to get rid of my local failure by upgrading gcc, g++ and libc6-dev [12:26] ok [12:26] jdstrand, I've been doing kernel snaps with all-root just fine [12:26] ok good [12:26] kyrofa, weird situation, abi breakage again? [12:26] it is only important for the system daemons to have their users around [12:26] hopefully I can just remove -all-root if it is an os snap and it will work. but I might have to disable the resquash for them [12:26] sergiusens, yeah not sure. Everything is broken! [12:28] kyrofa, tests seem to be running fine now for your stuff ;-) [12:30] mvo, ogra_: I added a card for -all-root and os snaps [12:30] thanks ! [12:31] will the ubuntu-device-flash available at http://people.canonical.com/~mvo/all-snaps/ubuntu-device-flash be packaged for ubuntu desktop 16.04 ? [12:33] ysionneau, there is a replacement in the works called ubuntu-image, though i was hoping mvo pushes the changes for bridging the time til thats ready if it takes longer [12:33] mvo, yeah still worth landing I say [12:33] * ogra_ isnt sure what the actual ETA for ubuntu-image is ... sergiusens knows though [12:35] ogra_, you didn't join the meeting yesterday; so you didn't figure it out [12:35] popey, I just noticed you tried to install the owncloud-pi gadget. Are you just looking for the owncloud snap? [12:35] ogra_, now you might [12:36] ogra_: ah ok [12:36] sergiusens, yeah, i totally missed while being busy releasing snaps to the store [12:42] ppisati, do you happen to have auboot binary for the rpi3 ? [12:42] * ogra_ is to lazy to compile one :P [12:52] hmmmm, so when snapcraft says 'architectures' when trying to build a gadget snap, it means? [12:53] kyrofa: yeah, i just thought I'd try installing that [12:53] popey, oh okay, so not really looking for anything in particular? [12:55] well, owncloud in particular, yeah :) [12:58] ysionneau, it means we have a bug :-) [12:58] ysionneau, can you log one? [12:59] popey, oh, just try `snappy install owncloud` [12:59] you mean can I file a bug report ? [12:59] ysionneau, yes please https://bugs.launchpad.net/snapcraft/+filebug [12:59] ysionneau, I will add it to the 2.6 work due tomorrow [12:59] kyrofa: on 16.04? [12:59] popey, indeed [12:59] I'm really not used to your bugtracker and I don't even have an account [12:59] but I guess it's the occasion [12:59] kyrofa: ok [12:59] let me try that... [13:01] popey, the gadget snap only exists for creating an image with the owncloud snap preinstalled [13:03] sergiusens, I swear every time I run the example tests I get one more failure. I'm up to 4 now [13:03] kyrofa, yeah I see the same [13:03] fgimenez, any help ^ [13:03] ? === chihchun is now known as chihchun_afk [13:04] The opencv failure is the same one I just got rid of locally by updating [13:04] I wonder if the others are similar [13:05] https://bugs.launchpad.net/snapcraft/+bug/1560481 [13:05] Launchpad bug 1560481 in Snapcraft "snapcraft prints 'architectures' when trying to build a gadget snap" [Undecided,New] [13:06] kyrofa: does it start automatically once installed? [13:07] popey, yes, though it takes a moment to generate mysql creds and stuff. Check the syslog [13:07] popey, you'll see apache waiting for creds, and mysql initializing [13:07] popey, on the rpi2, maybe a few moments ;) [13:08] it's been a while. [13:08] kyrofa, ogra_ jdstrand https://github.com/ubuntu-core/snapcraft/pull/391 [13:09] kyrofa: no apache or mysql processes running [13:09] sergiusens, looks good [13:09] popey, hmm, see errors in the syslog then? [13:09] [340124.599445] audit: type=1400 audit(1458651897.178:59): apparmor="DENIED" operation="capable" profile="owncloud.canonical_mysql_9.0.0ubuntu1" pid=10400 comm="mysqld" capability=24 capname="sys_resource" [13:10] kyrofa: ah, works on my edge pi2, but not my stable pi2 [13:10] guess this is expected? [13:10] popey, uh oh, no it's not [13:10] heh [13:10] popey, but I just released in an attempt to fix that [13:11] popey, I apparently failed [13:11] no worries [13:11] popey, can you give me the errors on the stable? [13:11] any particular logs? [13:12] http://paste.ubuntu.com/15471936/ [13:12] that's dmesg last bunch of lines [13:12] popey, just the syslog where they fired up. Those denials are normal [13:12] hope my bug ticket is clear enough sergiusens [13:14] ricmm: fyi, the youtube-streamer demo (I think it's yours?) spams the syslog every 5 seconds after install http://paste.ubuntu.com/15471942/ [13:18] ysionneau, no worries; I know what you are after :-) [13:19] ysionneau, in any case, gadget snaps in they future will be arch specific [13:19] so this is rather a remporary thing [13:19] ogra_: yep [13:19] i cant even get a purp out of mine (i finally built an aarch64 one) [13:19] ogra_: https://github.com/piso77/ubuntu-embedded/blob/master/boards/raspi3/rootfs/boot/firmware/uboot.bin [13:20] ogra_: and here is an ubuntu img, if you want to do some debug etc [13:20] ogra_: http://people.canonical.com/~ppisati/ubuntu_embedded/ubuntu-embedded-16.04-raspi3.img.xz [13:20] well, i hopefully only need a gadget snap :) [13:21] ogra_: +1 [13:21] bah+ [13:21] reading uboot.env [13:21] *** Warning - bad CRC, using default environment [13:21] missing our patch [13:22] ppisati, how exactly did you build it ? [13:23] ogra_: you mean the steps? [13:23] export ARCH=arm; export $(dpkg-architecture -aarmhf); export CROSS_COMPILE=arm-linux-gnueabihf- [13:23] make rpi_3_32b_defconfig [13:23] ppisati, oh, 32bit ? [13:23] make ... [13:24] ogra_: yep [13:24] ah, there is a special config ... thats my error [13:24] ogra_: my uboot binary is of some days ago [13:24] ogra_: i think back then there was no 64b option [13:25] rpi_3_defconfig defaults to 64 in swarrens branch [13:25] i just noticed that [13:26] grmpf [13:26] still doesnt boot [13:26] i suspect my trusty gcc is to old or something [13:26] ogra_: i'm on wily FWIW [13:39] popey: it's mine, and yeah, it will spam you until you configure it! [13:39] :( [13:39] popey: which is better than failing IMHO [13:39] we can maybe just make it print once [13:39] my poor sd card [13:39] it's been spamming for days [13:39] why did you install something without having it running? [13:40] to look at it [13:40] (hem, look at the logs, you have way more spammer ;)) [13:40] I can't believe you asked that question [13:40] but didn't configure it, so didn't look at it? :p [13:40] i wanted to see what was in the package, I didn't expect my syslog to be spammed for days as a result. "I'll configure that later" [13:40] "NO, YOU MUST CONFIGURE IT NOW!" [13:41] And I'll keep reminding you until your SD card wears out :) [13:41] popey: the issue is that if we didn't spam regularly, how would you know days after what to put in it? [13:41] Documentation [13:41] a README perhaps is the standard way [13:41] no way to expose the README in snappy [13:41] I can find it in the directory it's installed in [13:41] (forget people looking at the source git branch) [13:42] hum, I'm unsure [13:42] at least, it should print it once, ok [13:42] Thunderbird doesn't spam my syslog if I don't configure a mail account. [13:42] yeah, but configuring it is easier than snappy [13:42] Debateable :) [13:42] and you have a walkthrough [13:43] I don't believe that spamming the syslog is ever a right answer to "How should we tell the user that the app isn't configured" [13:43] ppisati, no go :( not sure why [13:43] * ogra_ just built on a wily machine [13:43] popey: remember you are not in the regular case, you are in the one installing something but not using it. However, I do agree, I can spam you once [13:44] popey: then, you go to answer people asking how to configure it :) [13:44] popey: mind opening a bug on the demo branch? [13:44] I'll do that on Friday [13:44] link? [13:44] http://github.com/ubuntu-core/demos [13:44] (I strongly disagree that people don't install and not configure stuff) [13:45] popey: I guess the real SD card issue is something else though, look at the logs you have, and yeah, we need ubuntu core to care about this [13:47] didrocks: done. [13:47] popey: link? [13:47] https://github.com/ubuntu-core/demos/issues/2 [13:48] ppisati, was there any script i need to run ? [13:48] ogra_: oh yeah, after yui build the uboot binary [13:48] ogra_: you need to run... [13:49] sergiusens: ack, thanks [13:49] i thought i remembered mkknlimage ... but cant find it [13:51] ogra_: mkknlimg [13:51] ogra_: https://github.com/raspberrypi/tools/blob/master/mkimage/mkknlimg [13:51] sergiusens: is there something I can do in my snap.yaml to temporary work around the issue to keep working? [13:52] ysionneau, use `snappy build` for now [13:52] ah, works, thanks! [13:53] ysionneau, note that will be going away soon, so move back to snapcraft once the bug is fixed [13:53] allright, noted, thanks! [13:58] sigh .. no go [13:58] sergiusens, no idea, it seems that 3 of the failures are consistent in each execution. we have a new test image fwiw, iirc that caused problems previously [14:00] how can I make a kernel snap? any example? [14:05] ysionneau, did you see the new Snapcraft release announcement? And sergiusens blog post? [14:07] kyrofa: was it on snappy-devel ? [14:07] I subscribed to this ML on the 10th of feb [14:08] ysionneau, snappy-app-devel [14:08] ah, I should subscribe to this one then [14:08] ysionneau, indeed! [14:08] Check the archives, sent out about 10 hours ago [14:08] done. [14:08] thx [14:09] https://lists.ubuntu.com/archives/snappy-app-devel/2016-March/000645.html o/ [14:09] awesome!! [14:11] ysionneau, bug reports welcome :-) [14:11] ah, you need to have snapcraft build your kernel, maybe I can hack something by creating a squashfs out of my already built kernel+modules+initrd [14:12] $ snapcraft --target-arch arm64 < !!!!! [14:12] --target-arch ? :) [14:12] so snapcraft is beginning to have some cross compilation notions ? [14:12] ppisati, i dont really get it ... do you do anything special ? [14:12] ppisati, i cant even get a burp out of the board with any uboot binary i build [14:13] ogra_: raspi3, right? [14:14] ppisati, yeah ... which compiler do you use ? just the gcc-5 eabihaf ? [14:14] *hf [14:16] ysionneau, only for kernels [14:17] but can any other plugin get the --target-arch value? [14:18] I'll have a look anyway [14:18] it would be cool if I can make my Alchemy plugin use that [14:18] so that I don't need to set the TARGET_CPU via environmnent variable [14:23] ogra_: don't you get to the uboot console? [14:23] ogra_: i mean, nothing [14:23] ogra_: zero output on the serial? [14:23] nothing [14:23] rightr [14:23] ogra_: oh wait [14:23] ogra_: is it the first time you use it? [14:23] if i copy yours in place it gets me the console but obviously the snappy patches are missing [14:24] yes [14:24] copying mine back in place gets me total silence on the serial [14:24] ok [14:24] let me try to rebuild one [14:25] mind throwing http://paste.ubuntu.com/15472180/ in ? [14:25] ppisati, you are using the swarren rpi_dev branch ? [14:25] yes [14:26] and only --dtok as option to mkknlimg ? [14:26] (ro also --283x= [14:26] --283x) [14:26] *or [14:27] (though it doesnt seem to make any difference here ) [14:27] my only idea is that we use different compilers or so [14:27] kyrofa, fwiw, I see opencv failing now too (on the builders) [14:28] sergiusens, urgh. Do the builders update first? [14:28] ysionneau, implement set_target_machine in your plugin [14:31] kyrofa, we can ask elopio like live now :-) [14:31] sergiusens, good idea [14:32] elopio, standup? [14:33] ogra_: i just compiled the tip of rpi-dev and uboot doesn't work here either [14:33] aha [14:35] do you knwo which commit you used ? [14:36] ogra_: ok [14:37] ogra_: git checkout -b rpi_dev_32bok f135e8f [14:37] ogra_: make rpi_3_32b_defconfig [14:37] ogra_: make ... [14:37] ./mkknlimg --dtok --ddtk --283x u-boot/u-boot.bin myuboot.bin [14:37] and it will work [14:37] just tested [14:38] either he broke 32bits or we are doing something wrong with the latest [14:38] but this will work [14:39] hmm git acts up [14:41] ogra@styx:~/all-snaps/rpi3/u-boot$ LC_ALL=C git checkout -b rpi_dev_32bok f135e8f [14:41] fatal: Cannot update paths and switch to branch 'rpi_dev_32bok' at the same time. [14:41] Did you intend to checkout 'f135e8f' which can not be resolved as commit? [14:42] ppisati, thats after a fresh: git clone git://github.com/swarren/u-boot.git [14:43] ogra_: i wonder if he deleted that commit, and the corresponding object is not present in the upstream repo anymore [14:43] :( [14:43] ogra_: i can give you a copy of my uboot tree [14:43] ogra_: until we find what's going on [14:43] yes please ... just tar it up or so [14:46] it's tarring... [14:53] swamped under hot tar ? [14:56] ogra_: http://people.canonical.com/~ppisati/u-boot.tgz [14:56] thx !!! [14:56] ogra_: today i got a new internet connection and a new router [14:56] cool ... how fast ? [14:56] ogra_: my firewall&c is a bit fucked up ATM [14:56] 100Mbit / 50Mbit [14:56] * ogra_ just switcvhed to 50/10 [14:56] d / u [14:57] the best i could get here [14:57] (after 10 years with 2/2 ...) [14:57] i guess you live far from a city [14:57] i live in the middle of a city :P [14:57] doh! [14:57] :) [14:57] germany is a third world country wrt internet speed [14:58] you must something really bad to your isp then [14:58] neeee [14:58] italy is worst [14:58] i justgot lucky [14:58] i have FTTH [14:58] we dont even have 2% fiber coverage afaik [14:58] and DSL max speed is around 100MBit ... but you have to be lucky [14:59] here vodafone is selling 300 / 30 fiber [14:59] but i didn't want to switch isp [14:59] (my wire could go up to 70 but then it would get shaky ... old cables ... ) [14:59] ppisati: I told you to do that... :-) [14:59] (now i'm on fastweb / swisscom) [14:59] Trevinho: you have fweb too? or vodafone? [14:59] ppisati: vodafone, but using telecom actually. [15:00] ppisati: I mean, they've not their fiber network here yet (building atm) [15:00] Trevinho: yeah, i was scared about downtime, etc [15:00] Trevinho: so i got an upgrade from my existing isp [15:00] ppisati: i've to say it's reaaaaally good in that terms. I was worreid, but it's not like their bad adsl [15:00] in most of germany the copper is still from the 70s ... i think thats the main prob [15:01] ogra_: i think pretty much everywhere copper sucks [15:01] ogra_: i must admit, having real fiber popping out of your wall is very nice :) [15:01] yeah [15:21] ppisati, ok, now i end up with FTD errors [15:21] *FDT [15:24] Hit any key to stop autoboot: 0 [15:24] libfdt fdt_check_header(): FDT_ERR_BADMAGIC [15:24] No FDT memory address configured. Please configure [15:24] the FDT address via "fdt addr

" command. [15:25] i got device_tree_address=0x02000000 in config.txt and also have my env set up to use fdt addr 0x02000000 [15:38] ppisati, do you have your uboot.env.in somewhere ? [15:45] sergiusens: it seems I need a snapcraft.cfg or something to use the kernel plugin? [15:45]