[04:19] I've just pulled the snappy-examples and built the hello-world, but get a bunch of errors and warnings about files which don't exist: http://paste.ubuntu.com/11488439/ Is that expected, or related to my env? [04:20] (it does build the snap, but unsure why I see those errors/warnings) [07:28] good morning [07:40] hi mvo, i'm trying the examples at https://code.launchpad.net/~snappy-dev/snappy-hub/snappy-examples [07:40] hey fgimenez, good morning [07:41] mvo, morning :) [07:41] mvo, on rolling/edge i'm getting the same aa_change_on_exec error you just reported for every snap [07:42] fgimenez: yeah, please use 15.04/edge for now, the rolling images are literally only ~30min old :) I enabled wily this morning on them and this was the first bug I noticed [07:42] fgimenez: I have not investigated further but it appears the apparmor profiles are not created [07:42] mvo, ok thx :) on 15.04/stable i'm getting this http://paste.ubuntu.com/11491522/ when trying to use the config hook, have you seen that before? [07:43] mvo, in the config-example snap [07:44] mvo, didn't try 15.04/edge though, doing it now [07:44] fgimenez: does the directory /var/lib/apps/config-example/1.0.6 exist? [07:44] fgimenez: it does not ring a bell, I will have to install/test in a VM to debug further [07:45] mvo, no, the one that exists includes the origin /var/lib/apps/config-example.sideload/1.0.6 [07:52] mvo, ogra_, asac: http://ubucon.de/2015 :) [07:52] woah, berlin [07:55] the cfp has started and they'd be happy if we could talk about all the new stuff we're doing and maybe do a workshop or two [07:57] fgimenez: aha, thanks. thats the bug then, the .sideload is not taken into account :/ [07:57] fgimenez: would you mind to file a short bug (this irc log is enough) so that its not forgoten? [07:57] fgimenez: should be easy to fix :) [07:57] mvo, of course thx :) [08:08] good morning. === devil is now known as Guest23013 === Guest23013 is now known as devil_ === erkules_ is now known as erkules [08:25] Morning all [08:33] morning [09:04] hi - i have trouble connecting to wi-fi with snappy ubuntu on a raspi2. i followed the wifi tutorial and installed the wireless-tools, etc. but if i connect the wifi dongle via usb, dmesg only shows that a new device was connected without loading any drivers or firmware. the interface wlan0 or similar also doesn't exist. how can i make it work? i had it running previously on an older image, so it should be working [09:06] Good morning all; happy Monday, and happy Say Something Nice Day! 😃 [10:43] mvo, beuno, hulp ! ... can one of you let chatroom through ? seems the review tools seem to be messed up ("ports is pbsolete in package.yaml") [10:43] *obsolete [10:52] ogra_: do you need this asap? [10:52] beowulf, mectors does i think ... i'm just trying an upload with changed ports directive though [11:07] mvo, could you take a look at https://myapps.developer.ubuntu.com/dev/click-apps/1526/ ? [11:08] i dont get why it spits out the two warnings (my snap overview page tells me 0.1-8 is published btw ... only the details page tells me it isnt) [11:18] can I freely change the hostname of a snappy core system? === dholbach_ is now known as dholbach [11:36] om26er, on it [11:36] er [11:36] ogra, on it [11:36] (who is not in the channel :)) [12:06] ogra_: hi, you can ignore those warnings for now, we need to stop generating those .snappy-systemd files, they are no longer needed [12:06] ogra_, I approved it [12:07] the fix is on the store side [12:07] I have it pending [12:07] beuno, ok, thanks ... if now the snap would actually work :/ [12:07] beuno, btw, where do i file store bugs ? [12:08] ogra_, https://bugs.launchpad.net/software-center-agent/ [12:08] * ogra_ has 5 pairs of identical mails for the last uploads ... no mention which version which mail belongs to [12:09] and my front apps page said the whoile time that 0.1-8 was accepted [12:09] * beuno nods [12:10] now why the heck does the chatroom not work anymore :/ === wesleyma` is now known as wesleymason [12:23] ogra_: maybe seccomp [12:23] sergiusens, hmm, i dont see any special messages about syscalls [12:24] what i see is: [12:24] Jun 1 12:23:01 localhost kernel: [ 2188.201850] audit: type=1400 audit(1433161381.662:146): apparmor="DENIED" operation="open" profile="chatroom.sideload_chatroom_0.1-8" name="/" pid=2356 comm="nodejs" requested_mask="r" denied_mask="r" fsuid=0 ouid=0 [12:24] but that has always been there [12:24] * ogra_ has an old install around that works fine on an old snappy [12:24] ogra_: why would you want to read / anyways ;-) [12:24] node does for whatever reason [12:24] ogra_: amd64 or armhf, I can try and install to see [12:25] but that never did any harm [12:25] this is currently amd64 [12:25] in kvm [12:25] mvo: any luck reproducing the issue wrt install not working? [12:25] my RPi is using a messed up image because i try to fix it :P [12:26] * ogra_ tries an older node-snapper tarball [12:29] Chipaca: no, but I didn't try further, sorry, I will try again in a little bit [12:30] ogra_: the one on the store works fine on 1.8 [12:30] ogra_: err [12:30] sergiusens: does 'snappy install' work for you in a vm? [12:30] ogra_: the one on the store, 1.8, works fine on bbb [12:30] ARGH !!! [12:31] sergiusens, because i'm the super idiot today [12:31] * ogra_ tested with firefox :P [12:31] ogra_: no binary for amd64? [12:31] (only works with chromium) [12:31] sure, it is multi arch [12:32] ogra_: ah, I tried on chrome fwiw [12:32] yeah, man, i wasted 2h on that :/ [12:34] * ogra_ adds that info to readme.md for next time :P [12:36] * ogra_ takes a break, phew ... [12:39] * Chipaca realises he hasn't updated in ages [12:45] hm, our builders are in a funny state [12:48] rsalveti, which ones ? images ? [12:48] normal lp builders [12:48] some are busy with the haskell uploads [12:48] but at our ppa all I get is 'start on' [12:48] no estimation at all [12:50] hmm [12:50] like https://launchpad.net/~snappy-dev/+archive/ubuntu/image/+build/7492026 [12:51] mvo: hm, ubuntu-snappy is also ftbfs on i386 for vivid [12:51] https://launchpadlibrarian.net/207810741/buildlog_ubuntu-vivid-i386.ubuntu-snappy_1.0.1-1%2B443~ubuntu15.04.1_BUILDING.txt.gz [12:51] yeah, i just saw the mail [12:51] rsalveti: I noticed but couldn't make sense of the failure yet [12:52] rsalveti, for the hanging build we should poke cjwatson i think [12:52] oh, thats a different error ... [12:52] * ogra_ had some with dependency issues on the weekend [12:52] ogra_: yeah, let's wait a bit more and ping him if still stuck [12:52] yup [12:53] yeah, why TestUpdateTimestamp would fail [12:54] wonder if a rebuild will make any difference [12:55] 2015-05-30 02:46:51 ERROR snappy logger.go:199 hello-app.potato failed to install: a package by that name is already installed [12:55] what about that one ? [12:55] wonder if that is testing the error itself [12:55] does it re-run the test in a loop or some such ? [12:55] since the test passed [12:55] ah [12:56] just a guess, didn't yet see the code :-) [12:56] let me open a bug for this guy [12:56] * ogra_ sighs ... i upgraded my laptop to vivid on the weekend ... got close to unusable now :((( [12:57] constantly to hot (80-90°C) [13:01] https://bugs.launchpad.net/snappy/+bug/1460650 [13:01] Ubuntu bug 1460650 in Snappy "(15.04) Rev 443 FTBFS on i386" [Undecided,New] [13:01] ogra_: any process in particular eating your cpu? [13:01] maybe new browsers [13:01] no, not at all [13:02] vivid is mostly fine for me [13:02] but starting chromium with no page open bumps the load to 12.00 ... while chromium only takes ~20% of one of the cpu cores (the rest is idle) [13:02] the frequesncy scaling seems pretty messed up === jgdxx is now known as jgdx [13:10] mvo, you should open up ownership of lp.net/ubuntu-core-launcher [13:10] Wanted to configure its bug tracker to show its bugs (right now I can't see bugs assigned to it because it's not set up) [13:10] nessita: what's the difference between price and prices, if the answer is, 'price' is legacy it's a good answer already :-) [13:11] so [13:11] on my system [13:11] mterry: I just ubuntu bug'ed it [13:11] snappy is bailing because it can't read /etc/localtime? [13:11] ah, no, it's before that [13:11] rats [13:11] sergiusens, correct! (price is legacy and is the price for the default currency( [13:11] * Chipaca carries on [13:11] __ [13:11] )) [13:11] sergiusens, you did...? [13:12] mvo: mterry: just put ubuntu-core-launcher under https://launchpad.net/~snappy-dev [13:12] rsalveti: excellent, thanks [13:13] i don't understand this failure :( [13:17] BAH ... i seem to have half a mobile desktop installed ... no wonder that laptop misbehaves ! [13:21] mvo: oh, sorry, I wanted to say that you could put it under ~snappy-dev :-) [13:21] only you can change it [13:23] nessita: I did.... nothing? [13:23] rsalveti: ups, sure [13:23] Chipaca: what's the issue? [13:23] Chipaca: btw, can you review my u-d-f change? [13:23] sergiusens: snappy internal unpack dies with "operation not supported" [13:24] and the strace shows no syscall failure [13:24] Chipaca: amd64 or bbb? [13:24] sergiusens: mad64 [13:24] rsalveti: updated, the project is now owned by ~snappy-dev [13:24] sergiusens, you meant mterry? [13:24] mvo: cool, thanks (mterry ^) [13:25] nessita: I did! [13:25] mvo, thanks! :) [13:25] mterry: ubuntu-bug ubuntu-core-launcher [13:25] but that will use the package bug [13:26] rsalveti: yes [13:28] the setuid/setgid fails [13:29] the setgid, in particular [13:29] can i say WAT? [13:29] Chipaca: uh, hope its not something silly as setuid and then setgid :/ in this particular revision [13:29] (i.e. order wrog) [13:29] no, it does setgid and then setuid [13:29] wrong even [13:29] ok [13:30] and the setgid fails [13:30] this is beyond weird, because the snappy we ship works [13:31] mvo, I also see that someone subscribed snappy-dev to snappy bugs. Maybe snapcraft and ubuntu-core-launcher should get the same treatment? [13:33] Chipaca: hmm, I face this problem when locally building snappy (I think) [13:34] mterry: I did that, yup, makes sense [13:34] rsalveti, thanks! [13:35] * Chipaca files bug #1460658 [13:35] bug 1460658 in Snappy "errors returned by syscall are not user-friendly, should not be shipped raw to the user" [Undecided,New] https://launchpad.net/bugs/1460658 [13:35] sergiusens: and how did you fix it? [13:41] Chipaca: I haven't, sorry [13:50] * Chipaca -> school run === joet is now known as josepht [14:08] nessita: can I use the 'id' for a package to get results for it? [14:09] nessita: as in using 2277 (id) instead of https://search.apps.ubuntu.com/api/v1/package/beagleblack [14:13] sergiusens, nopes, ID is retured by mistake,never rely on that [14:27] jdstrand: do you know why livecd-rootfs uses the "-M /var/cache/apparmor/.features" flag in ubuntu-touch? this file is not available in our chroot it seems [14:30] nessita: so the resource name is always the package_name.origin or alias, correct? [14:30] mvo: what is -M an argument to? [14:30] jdstrand: apparmor_parser -M (--features-file) I think [14:31] jdstrand: I'm not sure why its there for touch when they pre-build the /etc/apparmor.d/cache [14:32] ah, that is not in the manpage it seems [14:32] yes, I only found it with apparmor_parser -h and have no idea what its doing [14:33] mvo: iirc, if you specify a different --cache-loc (-L), then you need to have a .features file in it [14:33] ok [14:34] since there is /etc/apparmor.d/cache for system and /var/cache/apparmor for apps, we have a .features file in both [14:34] echo "I: precompiling click apparmor policies" [14:34] /sbin/apparmor_parser -M ${FEATURES} -Q --write-cache --cache-loc=/var/cache/apparmor/ `find /var/lib/apparmor/profiles/ -maxdepth 1 -type f -not -path '*/\.*'` [14:34] (thats from the code) [14:34] looking in the last image build log it seems to run fine without complaints [14:34] (touoch that is) [14:35] perhaps it is putting a .features file in place so that the compiler will use it instead of querying the kernel? [14:35] https://launchpadlibrarian.net/207811375/buildlog_ubuntu_wily_armhf_ubuntu-touch_BUILDING.txt.gz [14:35] i wouldnt know what would put that file there [14:35] I think rsalveti and/or jjohansen would know [14:36] I know that a .features file is put in place on the running system in each cache directory [14:36] the running non-livecd system that is [14:36] tedg, we should try to define a minimal set of snapcraft bits that we think we can move forward with this week [14:37] jdstrand, what puts it there ? the upstart job ? [14:37] and I know you can specify a different features file to adjust cache output [14:37] so you could compile caches for another kernel [14:38] there is definitely no code in livecd-rootfs putting that file in place [14:38] and i doubt it actually exists in the chroot after bootstrapping [14:39] sergiusens, yes [14:42] I'm not sure. I think the parser might [14:42] mterry, Yes, I haven't looked at the current code. I would like to get some stuff starting to work-ish [14:42] mterry, I think we can do that even if all the spec parts aren't perfected. [14:43] yes [14:43] If cache/.features is missing, create it if --write-cache [14:44] tedg, current code is an empty branch [14:45] I'm still going to defer to jjohansen [14:45] that makes it at least bugfree [14:45] mterry, tabula rasa [14:45] jdstrand, well, it doesnt seem to do any harm [14:45] (not sure what mvo's prob is with it) [14:51] sergiusens: mvo: it fails when compiled with golang 1.4 (!) [14:52] mvo: does this match the failures you were seeing in arm64? [14:52] or gogcc [14:53] maybe i just need to rebuild my go 1.4 [14:53] or switch to wily already [14:53] Chipaca: ohhhh, I have not tried that [14:53] Chipaca: in a meeting right now :/ so can't help just yet [14:53] mvo: is this the "15.04.1 release" meeting? [14:54] Chipaca: yes, I almost missed it fighting with gccgo [15:00] Chipaca: right, I'm using 1.4 here (locally installed) [15:01] sergiusens: vivid's 1.3 worked [15:02] sergiusens: locally installed 1.3 also works [15:25] I guess I won't be running snappy on my http://wiki.openwrt.org/toh/kingston/mlw221 any time soon [15:27] 16mb flash is going to be hard [15:28] pfft [15:29] just write a proper compression that allows pushing 4G into 16M ... [15:30] 4gb of /dev/zero surely [15:33] mvo: so, the manifest branch is the one causing the apparmor failures [15:44] Chipaca: the reduce-manifest one? [15:44] mvo: yaarp [15:44] ok [15:46] fgimenez: tomorrow I can escape at 11:30, london time. [15:46] mvo: still trying to understand why, but i'll get there :) [15:46] Where are the package lists for the default core image? [15:46] seed I guess? [15:47] Is it this? https://code.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/ubuntu-core.wily [15:51] elopio, fine, that is 10:30 my time. i'm already taking a look at the doc [15:51] elopio, ping me when you are ready [15:53] mvo, Chipaca: if I were to hazard a guess, I would think it has something to do with the 'namespace' changes in click.go [15:54] jdstrand: nope :) [15:54] oh, what is it? [15:57] fgimenez: great, thanks. [16:00] jdstrand: in revision 465 i switched some bits to use the yaml manifest instead of the click manifest [16:01] jdstrand: without noticing that things that appeared equal were not so [16:02] in my defence, in build we assign the click manifest's Hooks to be the yaml's Integration [16:03] but before doing that we do add some things to Integration that are not in the package.yaml [16:04] so, I *think* what we want is to grow Integration in the same way every time we load it [16:05] I need to check though :) [16:17] mvo: fyi, and this doesn't have anything to do with snappy internals, I have adjust the review tools to error if a snap uses 'integration' in its yaml [16:17] adjusted* [16:58] sergiusens: hey, I'm reflashing my bbb, is udf 0.21-1+174~ubuntu15.04.1 ok to use? [17:09] jdstrand: let me check [17:11] jdstrand: yes [17:18] So, I thought I'd go to the docs, didn't find my answer there either :-) [17:18] Where is the script/tool/etc that builds the core snap? [17:19] sergiusens: thanks [17:19] df [17:19] 1243 GB [17:20] 640k are enough for everyone ! [17:20] heh [17:31] * ogra_ upgrades ubuntu-device-flash [17:31] * jdstrand hrms [17:33] rsalveti, mvo: hello-dbus-app is broken on edge r72. is this the candidate for promotion? [17:33] rsalveti, mvo: (on bbb) [17:46] what would it take to port UbuntuCore to the Kindle Paperwhite [17:49] damjan, depends what bootloader it uses ... if the specs would work etc [17:50] hardware specs? [17:50] yes [17:51] you need 4G diskspace and the ability to partition it the right way... we currently only support u-boot on arm as bootloader ... [17:53] it's an 1Ghz imx 508 [17:53] 256MB ram [17:53] there's the 2GB and 4GB flash versions [17:53] that might eb a little low on ram ... but givenn that there is no UI support in any way yet, perhaps not [17:54] (i guess for plain cmdline 256M would work) [17:56] afaik it's u-boot [17:56] why does it matter btw? [17:57] because the automation for rollbacks lives in the bootloader [17:57] (if you upgrade and something goes wrong the system automatically rolls back to the last working setup) [18:00] it's uboot [18:43] rsalveti, mvo: fyi, there is a trello card to add hello-dbus-* to the self tests. if is in 'snappy core stackholders backlog' 'incoming propposals'. I just now added instructions on what to do (should be a very simple test) [18:43] rsalveti, mvo: s/if is/it is/ [19:34] Trying to do basic install of docker app on snappy. Installation completes w/o error however a simple `docker ps` fails: [19:34] ubuntu@localhost:~$ sudo snappy install docker Installing docker Starting download of docker.canonical 8.36 MB / 8.36 MB [======================================] 100.00 % 483.05 KB/s Done Name Date Version Developer ubuntu-core 2015-04-10 146 docker 2015-06-01 1.6.1.002 ubuntu@localhost:~$ docker ps aa-exec: ERROR: profile 'docker_docker_1.6.1.002' does not exist ubuntu@localhost:~$ [19:34] Ugh ... sorry about the formatting. [19:36] As you can see the complaint is regarding a missing "profile" file. I'm guessing I can get away w/ an empty profile file but ... what is the file name/location? [19:50] seb128: you around ? [19:51] kgunn, yes [19:56] jdstrand: ogra_: we only have the features file for touch [19:56] ubuntu-touch/includes.chroot/var/cache/apparmor/.features [19:57] so I'd guess the pre-cached apparmor stuff is kind of useless for core atm [19:58] jdstrand: broken on 15.04/edge? [20:00] rsalveti: that does sound broken [20:05] rsalveti: yes, r72 [20:06] rsalveti: I think it is surrounding the unshare in the launcher because I am getting a disconnected path apparmor denial [20:07] jdstrand: and thanks for updating the card, will move it around later today [20:07] but it is weird, because it is on run/ [20:07] but iirc it was working a few days ago [20:07] rsalveti: cool, thanks [20:07] I wonder what landed [20:07] now don't remember if that was indeed stable or edge [20:07] do we have a clear way to see what has changed? [20:08] not an easy way, no, ogra_ was working on that [20:08] we only got the manifest, but from a few images [20:08] I just reflashed my bbb today with edge and got this denial: audit(1433179950.388:19): apparmor="DENIED" operation="connect" info="Failed name lookup - disconnected path" error=-13 profile="hello-dbus-app.canonical_client_1.0.1" name="run/dbus/system_bus_socket" pid=1215 comm="dbus_message.ar" requested_mask="wr" denied_mask="wr" fsuid=1000 ouid=0 [20:09] notice 'name=run/...' [20:09] that should have a '/' in front of it [20:09] indeed [20:09] so, we knew we would need to adjust the apparmor template to use 'attach_disconnected' [20:09] and in fect, that is done in wily [20:10] fact* [20:10] but, I wasn't expecting this on 15.04, so I'm curious about the changes that went in [20:14] rsalveti: so, the launcher hasn't changed in vivid... [20:15] rsalveti: ah, 1.0.2~ppa1 is on the image [20:16] right [20:16] rsalveti: 'Set up a private mount namespace for /tmp' [20:16] so, this means we need a policy update [20:16] we use vivid + https://launchpad.net/~snappy-dev/+archive/ubuntu/image [20:17] yep, that is where I'm looking [20:17] * jdstrand sees the old unused libseccomp and deletes it [20:18] also noticed we're over the repo size limit in there [20:18] maybe because of gcc [20:18] oh yeah [20:19] rsalveti: so, at this point there is a choice, pull the launcher or have me do a policy update. I'm fine with doing an ubuntu-core-security update-- I can do it tomorrow. this was the other part of the sru I was talking about [20:19] rsalveti: but if we go that route, you should block the promotion on that [20:20] yeah [20:20] jdstrand: that same version is also available in wily (core-launcher) [20:20] so I guess we'd need a policy update for both [20:20] rsalveti: wily has 1.0.2 [20:21] wily already has the policy update [20:21] got it [20:21] 15.10.1 has: "add attach_disconnected for default policy in preparation of new /tmp [20:21] handling" [20:22] guess it'd be fine to get this in the ppa and then work on getting the official src in place [20:22] that is what I was thinking [20:23] cool [20:23] jdstrand: do we have a bug for this already? [20:23] just so I can put at the milestone [20:24] no [20:24] I'll file one [20:24] awesome, thanks [20:26] rsalveti: bug 1460810 [20:26] bug 1460810 in Snappy "1.0.2 launcher needs corresponding apparmor policy update" [Undecided,New] https://launchpad.net/bugs/1460810 [20:27] jdstrand: thanks [21:06] rsalveti: fyi, ubuntu-core-security - 15.04.12~ppa1 uploaded to the ppa [21:06] jdstrand: awesome, thanks for that [21:07] np [21:07] got the extra ppa space just in time [21:33] * Chipaca looks at sys/unix/syscall_linux_arm64.go [21:33] * Chipaca goes to bed [21:56] sergiusens, so, I have this item to sync with you, about snappy oauth-signing every single request to the index. Any pointers who should I ping about this? [22:20] nessita: for snappy? [22:28] Chipaca: is your snap removed from the store? [22:28] beowulf: unpublished [22:28] beowulf: should i republish? [22:28] people seemed to not like it being there [22:28] Chipaca: no, it exposed some wonderful errors :) [22:28] hee hee :) [22:29] Chipaca: so i was using it in webdm, then it was unpublished, then i tried to uninstall... [22:29] beowulf: webdm doesn't like things disappearing from the store? [22:29] Chipaca: no sir [22:29] *shock* [22:29] Chipaca: i'm wondering if the store should dissappear metadata, or should webdm handle it, or should i go to bed [22:30] i think the latter [22:30] jdstrand: is https://01.org/intel-kgt relevant to us? [22:30] beowulf: the magic 8-ball says "go to bed already" [23:56] Chipaca: beowulf this is a similar error to the offline case; I need to fix it [23:58] nessita: how soon is that timeline? I think you want to sync with beuno first on the whole concept for ssl/tls