/srv/irclogs.ubuntu.com/2015/06/24/#snappy.txt

=== utlemming` is now known as utlemming
=== fginther|lunch is now known as fginther
dholbachgood morning06:43
fgimenezgood morning07:01
=== blr_ is now known as blr
=== davidcalle_ is now known as davidcalle
Chipacamo'in08:32
Chipacasergiusens: when you're around, could use a hand reviewing the grub branches09:13
Chipacagetting "does not look like a tar" when using udf09:13
Chipacain other news, feeling much better \o/, and all my house smells delicious because i'm cooking up some chile :)09:14
Chipacaogra_: how're your teeth?09:14
ogra_Chipaca, i woke up with nearly no pain at all but a golfball on my jaw :/ ...09:16
Chipacaogra_: so you've gone to see a dentist09:16
Chipacaogra_: right?09:16
ogra_yeah, i fear i have to09:17
Chipacaogra_: dude, that's an abscess. you want that sorted asap.09:17
ogra_yes09:17
ogra_i fear there is not much to sort but get rid of the tooh09:18
Chipacathey can turn really nasty. don't google it, just go :)09:18
Chipacaogra_: forget the tooth, it's the ball of rotten pus inches from your brain you should be worrying about now :D09:18
ogra_if i open my mouth it is further away though :)09:19
ogra_(it is the lower jaw)09:19
Chipacaoh, that's alright then09:19
ogra_no, not, but not as dangerous :)09:19
Chipacatrue, true09:20
* Chipaca decides not to look up statistics09:20
ogra_lol09:22
sergiusensChipaca: which branch gives you that?11:00
sergiusensChipaca: does not look like a tar means an improper download maybe11:01
Chipacasergiusens: or maybe i'm using --device-part wrong11:01
Chipacasudo ~/canonical/goget-ubuntu-touch/src/launchpad.net/goget-ubuntu-touch/udf core rolling --device-part=generic-amd64_1.1_all.snap --channel edge -o wily.img --developer-mode11:01
sergiusensChipaca: you are11:02
sergiusensChipaca: that's supposed to be --oem generic-amd64_1.1_all.snap11:02
sergiusens:-P11:02
Chipacad'oh11:02
Chipacasergiusens: it worked!11:07
Chipacahere, let me pastebin it because it's loverly11:07
Chipacahttp://pastebin.ubuntu.com/11767162/11:07
Chipacasergiusens: ^11:07
Chipacasergiusens: (it was supposed to panic, right? :-p )11:08
Chipacasergiusens: and you've got a data race11:09
sergiusensChipaca: is that the last branch or one in between?11:09
Chipacasergiusens: the last one11:09
sergiusensChipaca: can you run godeps -u dependencies.tsv and rebuild?11:11
Chipacasergiusens: udf itself?11:11
sergiusensChipaca: yes11:11
Chipacasergiusens: but that'll fail, because i updated snappy manually to point at the branch you mention11:12
Chipacasergiusens: won't it?11:12
sergiusensChipaca: oh, don't use that snappy branch, it's irrelevant to u-d-f; it just needs to be on the image for updates on a running system11:13
sergiusensChipaca: it probably needs another merge from trunk11:13
sergiusensChipaca: and thanks, I will try with that branch in the meantime and figure out what's broken11:14
Chipacai'd pulled trunk and merged that branch, but anyway, rebuilding now11:14
sergiusensChipaca: yeah, this line doesn't feel right launchpad.net/snappy/partition.(*Partition).BootloaderDir(0xc208059700, 0x0, 0x0)11:15
sergiusensChipaca: which is most likely the developer mode stuff that landed11:16
Chipacasergiusens: i'll wait for you to fix then?11:16
sergiusensChipaca: I would want all the other developer-mode branches to land first; just godeps it11:17
sergiusensno need to wait11:17
Chipacak11:17
Chipacasergiusens: forwards, or back?11:17
Chipacaah, you mean the godeps you suggested above11:18
Chipacawhich i've already done11:18
Chipacabut still got the error11:18
Chipacabecause merge11:18
Chipacabzr revert'ed and trying again11:18
Chipacasergiusens: i have good news, and bad news :)11:21
Chipacathe good news is that i think it worked :)11:21
Chipacathe bad news is i build udf with -race11:21
Chipacaand it's not happy11:22
Chipacahttp://pastebin.ubuntu.com/11767220/11:22
sergiusensChipaca: oh thanks, now I have something fun to do today!11:34
Chipacasergiusens: you're welcome!11:57
* Chipaca turned his sarcasmeter off11:57
sergiusensChipaca: oh, I am indeed happy :)12:01
zygahey, who'd like to work with me to create a snap out of network manager12:28
zygaI tried to get started with deb2snap but it's not possible to handle n-m yet12:29
zygaand I kind of wonder what approach to use12:29
zygashould I built nm and its dependsncies myself12:30
zygaor try to reuse the existing packaging12:30
zygathen I wonder how to bridge network hardware to n-m12:30
zyga(and keep it dynamic so I can plug hardware later and still see it)12:30
zygaand how to start all the deamons that nm needs12:30
zygaand how to integrate that with udev on the device12:30
zygasergiusens, ogra_: ^^ any hints?12:31
verterokmorning12:39
zygaogra_: for networking, I remember what you said earlier about not depending on the core, for testing (wired) networking, should we bundle network tools (ifconfig, ping) or would you consider that to be an overkill12:49
ogra_zyga, no idea, thats an "architect" question ...12:51
zygaogra_: who should I ask?12:52
ogra_dunno12:52
ogra_lool, tvoss or ricmm perhaps12:52
zygalool: ^^12:52
ogra_i guess the low level bits will be in the image ... but something as high level as NM might have to become a framework or some such12:53
ogra_and it isnt even clear if we keep the if* tools or perhaps switch to systemd's network management daemon12:53
zygaogra_: hence I want to understand how we should rewrite our network tests12:54
loolzyga: you should bundle network tools, yup12:54
zygalool: during testing we'll also reconfigure networking12:55
loolthat's ok12:55
zygalool: will systemd undo/change any of our changes?12:55
loolzyga: not systemd12:55
loolzyga: it's using ifupdown ATM12:56
loolonly eth0 comes up automatically12:56
zygayes, I know but I recall that you want to move12:56
zygaand ifupdown is rather static12:56
zygabut systemd might be more proactive (I don't know)12:56
loolzyga: we aren't there yet though13:01
tomconteHi guys, do you know what is the status of Snappy images on Azure? I tried to create a few VMs and they all hang at 'Provisioning' time13:33
mvoutlemming: do you know if there is a issue with azure right now? see tomconte questions one line above13:35
mvotomconte: utlemming may know, but he is US timezone so the answer might be delayed13:35
tomconteNP thanks! the image I tried was something like Ubuntu-15.04-Snappy-core-amd64-edge-201506190757-94-en-us-30GB13:36
elopiohey sergiusens, I could use adt-run from trusty passing the ip of the testbed.13:47
elopiocan you paste the error you are getting please?13:47
sergiusenselopio: I got logged out, I can try again13:52
sergiusensChipaca: hey, forgot to send this your way https://code.launchpad.net/~sergiusens/goget-ubuntu-touch/downloadRace/+merge/26285113:56
sergiusensmvo: thanks for getting those MPs in13:57
mvosergiusens: hey, have you tested the lp:~sergiusens/livecd-rootfs/byeByeDebbie ? i.e. will apparmor work?13:57
mvosergiusens: yw!13:57
sergiusensmvo: only manually tested; I can do some more live testing than more than hello-world13:58
sergiusenswith more than*13:58
sergiusensmvo: btw, using the partition.new code in developer mode checks broke u-d-f :-P13:59
mvosergiusens: *cough* sorry for that, do you want me to do a branch?13:59
mvosergiusens: if hello-world works, thats probably good enough14:00
sergiusensmvo: no, no worries, it just caught me by surprise :-)14:00
elopiofgimenez: I changed the hangout name (once again) to not bother davmor214:02
elopioin case you are in the other one.14:02
fgimenezelopio, np, one minute14:03
davmor2elopio: now I'm going have to beat you for bothering me again,  Consider yourself beaten ;)14:04
elopiodavmor2: you asked for it14:04
sergiusensmvo: btw, ubuntu-snappy is failing to build14:05
jdstrandmvo: re byebyeDebbie-- we are talking about dpkg being gone right? /var/lib/dpkg/info/* will still be part of the readonly rootfs?14:21
sergiusensmattyw: I found your issue, well there's two actually, one is preventing the install14:37
sergiusensmattyw: if you introduce a security-override, aside from the apparmor entry you need to provide a seccomp one as well14:38
sergiusensmattyw: additionally, the apparmor profile has a read_path with a relative path, not sure why14:38
sergiusensmattyw: wrt seccomp, https://wiki.ubuntu.com/SecurityTeam/Specifications/SnappyConfinement#Seccomp14:40
mattywsergiusens, thanks for looking into it, I know very little about apparmor and seccomp - this is my introduction to them :)15:05
Chipacafgimenez: so!15:15
Chipacafgimenez: i am become help15:16
fgimenezChipaca, yes thx! :) let me paste the latests results...15:16
sergiusensmattyw: we all do, but luckily jdstrand knows some more, in any case the wiki has a simple yaml example for seccomp. You might also do an initial try on not requiring this apparmor override at all15:17
fgimenezChipaca, well, before of that, i'm creating the chroot with 'mk-sbuild --arch=armhf wily' is that ok?15:18
mattywsergiusens, juju trys to read /etc/os-release - I was getting app armour errors when trying to run it15:18
sergiusensmattyw: I see little sense in checking the release there as I guess you enable features depending on that entry and that probably means you toggle a lot of apt stuff which is just not there15:18
Chipacafgimenez: sounds about right :)15:18
mattywsergiusens, that's likely true, but I thought I'd try to get it working in snappy as is, then go back and re examine - it's a good excuse to learn about apparmor etc15:19
sergiusensChipaca: fgimenez are you going to do native qemu driven builds?15:19
Chipacafgimenez: what are we building, btw?15:19
sergiusensmattyw: right, just add the leading slash there ;-)15:20
mattywsergiusens, that's just a plain bug, thanks for spotting it :)15:21
fgimenezsergiusens, Chipaca i'm trying to build the snappy debs including the test ones using sbuild, not sure if that implies native qemu15:21
Chipacafgimenez: using sbuild, yes, it involves qemu15:21
sergiusenselopio: in any case ./run-checks is broken for me on trusty when adt is installed15:21
Chipacaat least, i think so15:22
Chipacafgimenez: or it might not! let me to the sbuild here too15:22
sergiusensChipaca: it does if doing native (qemu-static-$arch)15:22
sergiusenserr qemu-$arch-static15:23
Chipacasergiusens: “The auto-cross-builder runs in a fairly strict mode: in particular, it does not have qemu-user-static installed in either the host filesystem or the build chroots (previous incarnations did, but that is less useful for porting to new architectures)”15:23
sergiusensit may fail plenty15:23
fgimenezChipaca that might explain the error i'm finding, i've read somewhere about it being related to qemu... anyway, here it is: http://paste.ubuntu.com/11768227/15:24
fgimenezsergiusens, ^15:24
Chipacaah!15:24
Chipacayes15:24
Chipacayes, that's qemu15:24
Chipacafailing15:24
elopiosergiusens: I didn't try that directly because I have trusty in a kvm, and I can't create another inside as ./run-checks does.15:24
elopiowhat I did was to take our branch that passes an ip and port instead of installing the image in a vm.15:24
sergiusenselopio: oh, the bzr bd part of run-checks is what is bad15:25
elopioif your error is building the image, I can try trusty from a live cd. If it is on adt-run, I should be able to reproduce it this way.15:26
sergiusenselopio: it ignores what platform you are on so a 15.04 snapshot would be built with wily (or trusty or *) deps when it need be built with vivid ones15:26
Chipacafgimenez: so. You're needing the armhf package to install it in the image?15:26
elopiosergiusens: ah, that's because of your customized builder in bazaar conf.15:26
elopioor not.15:27
sergiusenselopio: it doesn't matter; if I remove that, it would build on trusty but it needs to have been built on a wily system15:27
fgimenezChipaca, yes, until we are able to get the image from the branch we should be able to compile them to install in the boards, what would be a good approach for that?15:27
elopiosergiusens: we can specify a --builder in bzr bd to select those options.15:28
Chipacafgimenez: which are the packages you need?15:28
elopioand we also copied your flag for receiving debsDir, so if you want to build the debs on your own, you just pass the dir.15:29
sergiusenselopio: right, if I build the debs on my own that means it can't really be part of run-checks, can it?15:29
sergiusenselopio: the only thing my bazaar.conf does is define a builder ;)15:30
sergiusenselopio: but it fails to understand -uc and -us15:30
fgimenezChipaca, we need the ubuntu-snappy packages from a branch, including the tests one to be installed on the boards http://bazaar.launchpad.net/~fgimenez/snappy/cross-compile-debs/view/head:/debian/control for instance15:31
sergiusenselopio: and it's something that needs to be built in to run-checks (passing --builder) so we all run the same thing15:31
elopiosergiusens: that's because the -uc -us are for debuild, and you are using sbuild.15:31
elopioif we always pass --builder debuild -- -uc -us it could work.15:31
elopiowhat I don't know is how to tell debuild to build for vivid or wily.15:32
sergiusenselopio: yeh, but I'm saying that using --builder debuild is the wrong thing to do ;-)15:32
sergiusenselopio: and the best way to convince you of that is to get you to run your main system on trusty ;-)15:32
fgimenezChipaca, ideally we should be able to build an image with them installed, so that we don't need to install them in a writable partition with dpkg (urgh), this is until we get there15:32
elopiotrusty, ugh... :)15:33
* Chipaca building a chroot by hand15:33
* Chipaca wonders whether making a snappy to build amrhf snaps would win him brownie points15:34
elopiosergiusens: we could always use --builder sbuild. But then we have to document on the README the schroots that we need.15:34
elopioI see no problem about that.15:34
sergiusenselopio: I'll create a branch for that as this basically blocks me from testing anything easily15:43
elopioI wonder if this would be easier with .dsc or .changes.15:43
Chipacawell15:52
Chipacaseems we have a broken wily build right now?15:52
Chipacasrc/launchpad.net/snappy/cmd/snappy/cmd_login.go:27:2: code in directory /tmp/cross-compile-debs/obj-x86_64-linux-gnu/src/code.google.com/p/go.crypto/ssh/terminal expects import "golang.org/x/crypto/ssh/terminal"15:52
elopiosergiusens: thanks.15:55
elopiofgimenez: either the test finishes before the reboot and it never happens, or it receives the segkill.15:56
elopiothe only option I see is to do the reboot outside of the test binary.15:56
elopiowe can do that with a wrapper script, and leaving a file /tmp/needs-reboot or something.15:56
fgimenezelopio, sounds good, it can play well with adt-run16:05
fgimenezChipaca, i saw something similar, in my case it was looking for dependencies in /usr/share/gocode (which belongs to root) instead of the defined $GOPATH, don't know if it might be related16:08
Chipacafgimenez: i got to the point where it got building, but then failed on some bits that don't work for cross-compiling. i'll retry in a bit16:09
Chipacafgimenez: is it bad if i skip the package build tests?16:09
Chipacathat is, i don't run the tests during package build of this package16:09
fgimenezChipaca, i think so, elopio? ^16:10
elopioChipaca: not if you are running from ./run-checks.16:11
elopiowhen you ./run-checks, it will run the tests, and then run them again during the bzr bd.16:12
Chipacayep16:12
Chipacaexcept i was hoping to not have to get the tests to pass inside a bodged up chroot :)16:12
fgimenezChipaca, elopio leaving, will catch up tomorrow16:21
Chipacasergiusens: is there a way to make dh_golang specify -ldflags?16:48
sergiusensChipaca: no, I wish there were16:55
sergiusensChipaca: I also wanted to do some neat tricks with package building for autogen'ed stuff16:55
Chipacasergiusens: this will make my evil plan of compiling snappy with musl a lot harder16:59
sergiusensChipaca: heh, if dh_golang had prov and build in different steps it would be rather easy though17:03
sergiusensChipaca: btw https://code.launchpad.net/~sergiusens/snappy/seccomp/+merge/26288017:03
sergiusensoh snap, my fingers copied an extra line there :-P17:04
sergiusensthere we go17:05
cyphermoxelopio : just checking, do you know if today's snappy personal EFI image booting properly now?17:10
elopiocyphermox: I didn't know it wasn't booting properly before. I haven't touched personal yet.17:11
cyphermoxAh, OK. sergiusens suggested you'd know :-)17:11
Chipacasergiusens: +117:12
Chipacaand i'm off to have glorious, glorious dinner17:12
* Chipaca 's been smelling it all day17:12
elopioChipaca: mmm, enjoy.17:13
sergiusenscyphermox: oh, you meant personal?17:13
sergiusenscyphermox: we don't qa personal, that's on the personal team17:13
elopiocyphermox: I can give it a try. I think I just need to build ubuntu-device-flash from source.17:13
elopios/source/trunk.17:13
sergiusenselopio: it's in wily now fwiw17:13
elopiosergiusens: ah, that's good. Thanks.17:14
sergiusenselopio: https://launchpad.net/ubuntu/+source/goget-ubuntu-touch/0.25-0ubuntu117:14
elopiosergiusens: what does that personal-oem contains?17:15
cyphermoxsergiusens : sorry, I wasn't clear enough, I mean the same image that seb reported didn't boot successfully in UEFI17:15
cyphermoxI can just as well test it myself later if having it working isn't an omg-fire :-)17:16
sergiusenscyphermox: ok, so I have no efi machine, it's just a seb thing :-)17:18
cyphermoxOK. Well it can be tested with qemu and OVMF, but I'll play with it tonight :-) I don't want to distract you from more important things17:19
elopiocyphermox: that sounds interesting, and something we should automate somewhere.17:22
elopioI'll make a card to test with https://wiki.ubuntu.com/UEFI/OVMF17:22
elopiosergiusens: if we run the selftests on an image with personal-oem and a qemu with ovmf, do you think we should put that in lp:snappy ?17:24
cyphermoxElopio : I can share with you the qemu options to use17:26
elopiocyphermox: yes, please.17:27
elopiohttps://trello.com/c/I1BSZ6iV/130-run-tests-for-snappy-personal-and-uefi-boot17:27
tomconteutlemming: Hi, do you know what is the status of Snappy images on Azure? I tried to create a few VMs and they all hang at 'Provisioning' time17:32
utlemmingtomconte: which image?17:32
tomconteutlemming: the image I tried was something like Ubuntu-15.04-Snappy-core-amd64-edge-201506190757-94-en-us-30GB17:33
utlemmingtomconte: there is a bug with Snappy on Azure that needs to be fixed, that we have been trying to clear. I'm waiting for feedback from Microsoft.17:33
utlemmingtomconte: but, yeah, I am waiting that situation. The issue is that the /writable partition is not being seen at boot.17:34
utlemmings/waiting/watching/17:34
tomconteutlemming: I can try to help, could you forward me the issue or add me to the email loop ?17:36
tomconteutlemming: I am from Microsoft :)17:36
utlemmingtomconte: yeah, let me see if I can dredge up the bug report on it17:38
sergiusensChipaca: do you have a minute for a hangout?17:57
Chipacasergiusens: now i do18:16
Chipacaelopio: https://goo.gl/photos/73DdNVzZYeEa1ZMb918:17
rsalvetisergiusens: elopio: so let's target the next stable release for july 10, since next week we're going to have a new kernel18:20
rsalvetiso target for first RC could be july 618:20
rsalvetielopio: would probably be good to do another round of bug triaging as well, so we can target possible fixes or backports we can still work on during next week18:22
rsalvetihttps://launchpad.net/snappy/+milestone/15.04.2 with the date18:25
Chipacasergiusens: am going to be around for a while, but am officially breaking out the beer in a few18:30
sergiusensChipaca: ok, just wanted to talk about _$version and the future18:33
sergiusensChipaca: https://plus.google.com/hangouts/_/canonical.com/days_of_future_past18:36
elopiorsalveti: great, thanks.19:26
sergiusensrsalveti: the problem with google docs instead of VCS is that it's hard for people to spot the updates ;-)19:50
marcoceppihello! I have a raspberrypi2 and I'm trying to figure out how to flash it...20:25
marcoceppiI found this: http://people.canonical.com/~platform/snappy/raspberrypi2/ but I have no idea what I'm doing20:26
sergiusensmarcoceppi: http://developer.ubuntu.com/en/snappy/start/#snappy-raspi220:29
sergiusensmarcoceppi: with the images you found there20:29
marcoceppisergiusens: thanks!20:29
sergiusensalecu: kyrofa do you know how eaasy or hard it is to see changes in a google doc? I was udating here and there20:30
alecusergiusens: we can only see the revision history if we have Edit permission20:32
alecusergiusens: "File->See revision history..."20:33
sergiusensalecu: then yiu shall get them!20:33
sergiusensalecu: there!20:34
alecusergiusens: the history view is not great, but it aggregates changes that happened close in time, so it's possible to understand what was changed.20:34
alecusergiusens: I'll try reloading, thanks.20:34
alecusergiusens: looks good, thanks.20:35
rsalvetisergiusens: Chipaca: elopio: what are we missing for https://code.launchpad.net/~fgimenez/snappy/go-functional-tests/+merge/261748 ?22:22
elopiorsalveti: one review.22:23
rsalvetisergiusens: Chipaca: can either of you guys take another round of review for that branch ^?22:23
Chipacarsalveti: sure22:24
Chipacarsalveti: in the morning :)22:24
rsalvetiChipaca: sure :-)22:24
Chipacaelopio: for future reference, it's likely i would have reviewed it by now if it weren't asking for a review from mvo explicitly22:25
Chipacaor more likely, at least :)22:25
elopioChipaca: I added mvo just this morning, thinking that a direct ping to somebody would speed things up ;)22:26
Chipacaelopio: orly?22:26
Chipacaelopio: my apologies then!22:26
marcoceppigetting an error when trying to boot on the ubuntu matchbox rpi222:28
marcoceppisystemd failes to load default taget: no such file or directory,?22:28
marcoceppihttps://lh3.googleusercontent.com/kbsG4MKVxq6SbdiCHgKnf1ttfZcSdBUf-f4RAYn1W3dF=w1342-h993-no22:31

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