[03:12] Yesterday, I Downloaded Snappy from "http://releases.ubuntu.com/15.04/ubuntu-15.04-snappy-armhf-bbb.img.xz"for BBB [03:12] But the system didn't boot [03:12] I checked the "system-boot partition, the "uenv.txt" was empty [03:13] "system-boot" partition [03:14] And there was a install.yaml! It was not available in the version 1! [03:14] Well, do you know how can I boot the new snappy on my BBB? [03:16] Should I have to copy the uenv.txt content from version 1 to this new version? or the boot mechanism is totally changed in new version? [03:18] Here is the uenv.txt in version 1, 2 and 3 -> "http://paste.ubuntu.com/12720512/" [03:18] It's same [05:49] No solution for boot problem?!!! [06:11] Rlyeh, afaik, uEnv.txt should be empty. There is snappy-system.txt with all required variables [06:17] It's not booting :( [06:23] BBB jumps to debian on emmc [06:33] Rlyeh: did you hold down S2 before connecting power to the BBB? [06:48] Buongiorno :) [07:12] good morning [07:24] Good morning [07:28] good morning [08:08] good morning snappy [08:09] biezpal: it is normal that uEnv.txt is empty. Snappy uses uboot.env for booting [08:10] biezpal: Do you have a serial console, to check what u-boot is doing? [08:11] biezpal: ah sorry, it was Rlyeh who asked that question [08:14] longsleep, np :) [08:22] well, Ryleh most likely didnt press S2 [08:43] ogra_: i should probably get a BBB for testing - S2 is a button? [08:44] yeah ... also called "user button" [08:44] ok - and that controls what to boot or why does one have to press it? [08:44] you have to hold it down for the first time you want to boot from external media (subsequent boots from SD dont need it) [08:45] ahh - the bbb only boots from nand by default - i get it thanks [08:45] well, eMMC [08:45] is the eMMC removable ? [08:45] nope [08:46] i see - can snappy also boot when put on the eMMC then? [08:46] never tried :) [08:46] longsleep: you can nuke the MLO on eMMC, then it will default to SD [08:46] tbr: right, but would be nice to avoid using an external media or have the sd available for storage [08:46] buut given we boot based on labels i wouldnt see why not [08:47] as long as you can name the partitions correctly on the MMC [08:47] it works fine for the odroid, though the eMCC is removable there and we do not use it because of the cost [08:47] ogra_: I offered a while ago to look into the install script that debian uses and how that could be adapted, but then some handwaving ensued [08:48] how do you flash the eMMC - is there some boot loader / fastboot thing? [08:48] i mean if you cannot put it into your pc [08:48] longsleep: you could flash it over USB or UART, but usually you just boot from an SD and write from the booted system to the eMMC [08:49] there is work going on to create a recovery img we ship in all installs ... [08:49] that would have a dd like installer i guess [08:49] and factory reset etc [08:50] yeah that would be nice [08:53] tbr: i just read the docs about it, thanks [08:54] ogra_: do you have an idea what happens if you have partitions with the same label on eMMC and SD ? [08:54] * longsleep wonders which one is used [08:54] nope [08:54] havent tried [08:54] Good morning all; happy Friday, and happy Curious Events Day! 😃 [08:56] https://github.com/RobertCNelson/tools/tree/master/scripts - if someone wants to take a look how to adapt === vrruiz_ is now known as rvr [09:57] elopio: fgimenez: you around? [10:18] Chipaca, yep [10:18] fgimenez: i have a branch up for review that probably breaks stuff [10:19] fgimenez: could you / would you take a look? [10:19] fgimenez: https://code.launchpad.net/~chipaca/snappy/auth/+merge/273684 [10:19] Chipaca, sure, which one? [10:19] Chipaca, ok thanks :) [10:19] fgimenez: basically all non-GET rest api tests will fail [10:20] fgimenez: or should :D [10:20] need to add something to first boot to set a default password [10:20] Chipaca, ok, thx i'll check :) we are currently testing post to packages [10:20] neat [11:19] Chipaca, it fails \o/ [11:19] heh [11:20] fgimenez: thanks [11:20] now we need to decide how to un-fail it :) [11:20] Chipaca, ok, the tests need some adjustments though, it was passing at first :) [11:20] fgimenez: oh? how? [11:21] Chipaca, they was using /usr/bin/snapd instead of the compiled one [11:21] lel [11:23] Chipaca, also, we were checking for a 405 method not allowed response in unhandled verbs [11:23] that is correct [11:23] but now you'd get a 401 before that [11:23] Chipaca, yes, we need to change that in the test [11:23] you'd still want to check for the 405 once you're sending in credentials [11:23] Chipaca, ah, ok, then that's ok [12:17] ogra_: you around perchance? [12:17] no, not perchance ... [12:17] ... just regular [12:18] :) [12:18] :) [12:18] ogra_: so, REST API now needs creds to do stuff [12:18] whee [12:18] ogra_: including things like, um, setting up creds [12:18] lol [12:18] (nothing of this is landed yet ;) [12:19] so i guess we should have some snappy config way to set this to a sane default (i.e. if --developer-mode enable ubuntu/ubuntu) [12:19] that, and/or "snappy passwd" [12:20] ah, thats new ? [12:20] that's non-existent [12:20] that's what i'm asking :) [12:20] can we also enforce updating of the pw ? [12:20] ogra_: you mean, pw age policy? [12:21] so that we could ship with a default on the images but enforce the update on first login via the REST api ? [12:21] yeah, something like the pw age policy [12:21] but if i dont have ssh enabled i need to be able to get in once to set a new one [12:22] that's a good point, we probably want that [12:22] but it's a chunk of work :) [12:22] yeah, i guess [12:23] and not sure of the details of it [12:24] ogra_: snappy config such that the oem can set default pw was what i was thinking, with it defaulting to ubuntu/ubuntu if in dev mode [12:24] ogra_: and i didn't know if we'd also need a snapp passwd thing [12:24] ogra_: password aging, and password policy in general, wasn't in what i'd been thinking of doing [12:24] well, we dont use --developer-mode on the official images we offer [12:24] only --enable-ssh [12:25] so the cmd is a must [12:26] with a nice-to-have of accepting config (just to set passwd in core) if no passwd set [12:26] also do we have a way to have passwords hidden in the snappy config output yet ? [12:27] so only *** is shown [12:27] ogra_: i was thinking of doing it for coreconfig [12:27] yeah [12:28] though with a fixed lenght string so the actual PW lenght isnt guessable [12:28] ogra_: was thinking of just replacing the value in the yaml with "***pasword***" or some such placeholder text, on the way out [12:28] yeah [12:28] ogra_: which is something you could do with bip too :) [12:28] just to indicate there is one set [12:28] (was it bip?) [12:29] yeah [12:29] the package is called ircproxy :) [12:29] but it uses bip [12:30] ogra_: another way would be to use a special yaml thing for passwords [12:30] i could dig into that if it'd be useful [12:31] we should ask jdstrand for input here [12:31] after all thats a security thing :) [13:28] ogra_: echo "config: {ubuntu-core: {password: password}}" | snappy config ubuntu-core -- - [13:28] \o/ [13:29] :) [13:30] Hi Chipaca, any news from security guys for Bug #1496319? :) [13:30] bug 1496319 in Snappy "Could not create symlink to hw device with udev rules" [Undecided,New] https://launchpad.net/bugs/1496319 [13:30] clobrano: no [13:30] clobrano: but i hear they're backlogged atm [13:31] and most of them are in the US.... [13:31] where nobody works today... celebrating that guy who got lost [13:33] :D [14:03] Chipaca: we are we needed for 1496319? === davidcalle_ is now known as davidcalle [14:11] did anyone look into snappy images for scaleway yet? [14:17] beuno: fyi, I tried to accept docker in the review: OOPS ID: OOPS-3587e98ff8044c0db2885414f81ffd46 [14:17] kickinz1: ^ [14:18] beuno: I don't know how to proceed and am afraid to try to approve again [14:18] beuno: it is ok to accept (kickinz1 will fix 'architectures' for next time and the security policy changes are ok) [14:19] it appears to have been accepted [14:20] (based on the website) [14:31] jdstrand, I accepted it as well [14:31] maybe we fought over accepting it [14:31] jdstrand, I'll look into it, thanks [14:43] maybe [14:51] ogra_: That's Monday actually, US is in today :-) [14:51] (well I imagine some people took vacation) [14:57] tedg, oh, not today ? [14:57] damn, mixed that up [14:58] tbr, i dont think anyone has started on that [14:58] k [14:58] it's interesting because it's cloud, but it's ARMv7 [14:58] yeah [15:22] fgimenez: any idea why umount would pin a cpu at 100% at the start of adt-run? [15:23] 32223 root 20 0 25392 1688 1512 R 1 100. 0.0 8:52.81 umount /tmp/diskimage185973715/system [15:23] like so ^ [15:24] Chipaca, nope atm, how do you run it? [15:24] fgimenez: GOPATH=~/canonical/snappy ./run-checks [15:26] Chipaca, never saw that before, from which branch are you running it? [15:27] fgimenez: my auth thing [15:28] Chipaca, didn't try run-checks, go run _integration-tests/main.go goes fine here, i'll check run-checks [15:28] it's u-d-f hanging somehow [15:28] sergiusens: you seen that? [15:29] Chipaca, fgimenez if u-d-f was rebuilt, maybe check all the FDs are closed [15:29] lsof [15:29] it shouldn't cause a spin of the cpu [15:31] oh [15:31] no space left on device [15:31] that might cause it [15:31] fgimenez: could we use a link to the icon, instead of copying it into every snap? [15:31] does that work? I'll give it a try. [15:33] elopio, it would be very useful, let me know how it goes [15:41] fgimenez: I did this service test before I realized I wanted to copy your local snaps style. [15:41] I regret that because now every other branch depends on it. [15:41] are you ok if I land it, and then make another branch to move the helloDBus tests to examples, and add a local snap instead? [15:43] elopio, sure, np [15:44] fgimenez: rebooted because my /tmp got into a mess, and now i get past the unmounting, but am still at the "ssh connection failed" bit [15:44] i've never gotten past this stage [15:44] how do i get it past this stage? [15:46] Chipaca, which version of autopkgtest are you using? [15:46] fgimenez: 3.17.3 [15:47] Chipaca, same here, what if you execute go run _integration-tests/main.go -release 15.04 ? [15:47] Chipaca: did your udf fail? I get this every other time: https://bugs.launchpad.net/ubuntu/+source/goget-ubuntu-touch/+bug/1496484 [15:47] Launchpad bug 1496484 in goget-ubuntu-touch (Ubuntu) "device-mapper: remove ioctl on loop0p5 failed: Device or resource busy" [Undecided,Confirmed] [15:48] elopio: when it failed, i aborted, cleared loop0p5, and redid [15:49] fgimenez: doing that now [15:49] tedg: how do I make a snappy lxc? [15:53] fgimenez: getting the same “adt-virt-ssh: WARNING: ssh connection failed. Retrying in 3 seconds...” [15:53] fgimenez: except more than 3 seconds pass between each one [15:53] and the qemu is doing *something* [15:57] Chipaca, you should be able to access the image with kvm -m 512 -redir :8022::22 /tmp/snappy-test/image/snappy-rolling-edge-latest.img [15:57] Chipaca, does it boot ok? [15:58] well, 15.04 not rolling because you just told me to -release 15.04 :) [15:58] * Chipaca tries it [15:58] Chipaca, yes, sorry :) [15:59] it's not booting [16:00] Chipaca, maybe udf doesn't create it properly [16:01] is that 204 ? [16:01] (that is exactly what i wanted to have tested) [16:01] could be the grub changes, but these only landed in 204 [16:03] this is 15.04 [16:03] not sure the revno [16:03] redoing from the start [16:03] u-d-f should have printed it [16:04] ogra_, 15.04 204 is working fine here [16:04] phew ! [16:04] * ogra_ hugs fgimenez [16:04] thanks ! [16:05] that saves my weekend :) [16:05] ogra_, yw :) [16:06] * Chipaca continues to work to ruin ogra_'s weekend [16:06] haha [16:06] * Chipaca likes the feeling of power [16:06] my heating breaking down already is ahead of you [16:06] * ogra_ hopes it is only the oil filter being dirty and not the pump [16:07] oil filter? do you use a diesel engine to power the pump? [16:07] yeah, but its a clean VW one [16:07] :P [16:08] ogra_: damn you beat me to it :) [16:08] (i think 60-70% of german heatings still burn oil) [16:08] fgimenez: retried with rolling, still no luck with the tests, but the image boots [16:08] its a central heating but a oli furnace [16:08] ... that powers it [16:09] fgimenez: and i can ssh in and all [16:09] ogra_: yes you and america are the reason the cost of diesel rockets in the winter [16:09] i dont really need the heating, it is warm enough still, but it also does warm water [16:09] ogra_: leaves more russian gas for the uk \o/ [16:09] lol [16:09] i wish we had gas in my street ... [16:10] but like cable the neighborhood wasnt willing to pay for it [16:11] (tow streets down they have gas and 200MBit cable) [16:11] Chipaca, what if, with the image booted with kvm, you run go run _integration-tests/main.go -ip 127.0.0.1 -port 8022 ? [16:12] STUFF IS HAPPENING!!!!1!!!one! [16:12] yay [16:12] Chipaca, \o/ [16:14] this bypass the image creation and booting part, anyway something is wrong, that should be done by the runner [16:22] stgraber: I thought there was a snappy image for LXD, but I can't seem to find it. Am I remembering wrong or just suck at Google? :-) [16:23] tedg, there is an lxd snap in the store [16:24] ogra_: Yeah, looking the other way around [16:26] fgimenez: ohhh... [16:26] fgimenez: after a number of reboots and stuff, this: [16:26] Rebooting... [16:26] bash: line 1: 1087 Killed bash -ec './_integration-tests/reboot-wrapper integration.test' 2> >(tee -a /tmp/adt-run.rEvouJ/command1-stderr >&2) > >(tee -a /tmp/adt-run.rEvouJ/command1-stdout) [16:26] fgimenez: and now it's stuck waiting for ssh again [16:26] (i can log into the system just fine) [16:26] ah! but no ipv4 address! [16:28] * Chipaca runs dhclient by hand [16:28] Chipaca, you found it after all! i thought i was mistaken :) it must be the same think that elopio reported for bbb [16:35] leaving, nice weekend everyone o/ [16:54] ogra_: elopio: have either of you ever seen a package decide it's sideloaded on update? === fginther` is now known as fginther === lborda is now known as lborda-sprint [17:19] Chipaca: nop. [17:20] Chipaca: oh, I forgot. I did this because it was hard to find the test that was failing for you a couple of weeks ago [17:20] https://code.launchpad.net/~elopio/snappy/results_on_error/+merge/273013 [17:21] What is the best approach to add snapcraft plugins, adding an example how to use it? [17:21] not sure if there's a better way to do it. Can you review please? [17:21] tedg, until recently there was no snappy image for lxd due to all the security architecture around it [17:22] recently though, apparamor gained the possibility of profiles within profiles which was one of the blockers [17:22] * sergiusens knows he is using the terminology all wrong [17:26] wait, what? [17:26] I'm confused by the 'all wrong terminology' [17:27] oh, snappy in lxd? [17:27] right, so, apparmor doesn't yet have namespace stacking support [17:28] which means that the host and the guest can't have unmodified profiles [17:29] so it makes running snappy in a container harder than you'd want (should be able to if you disable apparmor for lxd on the host) [17:29] the namespace stacking work has not landed yet, but it is coming. we are targeting 16.04 [17:37] Ah, okay, so elopio I was wrong about that ^ [17:38] Apparently we can't do the tests in LXD yet. [18:17] Chipaca, never [18:31] longsleep, I don't follow, or if I do, just add a small project in the examples directory; to make sure it doesn't break even add a smaller project in integration_tests [18:55] beuno, you are being missed in a certain "-external" channel [19:45] ogra_: fyi, I filed those three snappy bugs we talked about (nameservers, ntp and rsyslog) [21:25] jdstrand, I think a little bird told me you were going to get the reviewer tools updated in ppa:snappy-dev/tools for vivid and trusty, is that correct? [21:25] yes [21:26] but I need to first update them for the things we agreed to :) [21:26] I've started that [22:16] sergiusens: ok, fix committed to trunk. the store will pick it up next week [22:16] sergiusens: I have one other thing to add then I'll push to the ppa [22:16] that will happen next week [22:17] sergiusens: unless it is critical that you have it in the ppa sooner. let me know and I'll push what I have [22:20] jdstrand, it's fine; just noticed an MP where snappy build re enabled the review tools so didn't want have havoc when the tools ppa got updated [22:24] sergiusens: right, don't push that yet. feel free to add me as a reviewer [22:25] sergiusens: the one thing that is left is actually so that can be turned on [22:28] jdstrand, I noticed it landed while I was out; but it hasn't made it to the general population yet ;) === Guest87130 is now known as balloons [23:08] sergiusens: ok, I uploaded review tools 0.34 to the tools-proposed ppa [23:09] * jdstrand -> weekend