=== chihchun_afk is now known as chihchun [07:04] good morning === clobrano is now known as c-lob [07:15] good morning === c-lob is now known as clobrano [09:25] Good morning all; happy Monday and happy Watermelon Day! 😃 [09:26] yummy ! [09:26] matches the weather :) [10:26] ppisati, is https://github.com/piso77/ubuntu-embedded/tree/master/boards/raspy2/bootloaders just a cloned snapshot of https://github.com/raspberrypi/firmware.git ? or did you compile anything in that dir ? [10:27] * ogra_ would like to use upstream for the oem snap, but if that needs modification i'll fall back to yours) [10:27] ogra_: no compilation [10:27] ogra_: it's a copy of the binary stuff that was shipped with the original raspi2 img long ago [10:28] ogra_: let me check one thing [10:28] k, thx ... they seem to set a kernel version inside their dtb's there [10:28] (which points to 4.0.9) [10:28] ogra_: i didn't update mine [10:29] ogra_: so they probably update it [10:29] yes, they do [10:29] since we need to use their dtb's for the initialization i wonder if that could cause issues [10:30] rsalveti: I guess we might want to escalate https://bugs.launchpad.net/bugs/1470727 to Critical as it prevents creating images of rolling? or is there another way to get to latest rolling edge? [10:30] Launchpad bug 1470727 in Snappy "ubuntu-device-flash touch fails with: failed to find user uid/gid" [Undecided,Confirmed] [10:31] lool, that should be fixed in the proposed PPA asince a while [10:31] (I tried updating form oldest working image but that got me into a broken state -- empty GRUB menu) [10:31] ogra_: which one is that? [10:31] I have the snappy-dev tools PPA only [10:31] https://launchpad.net/~snappy-dev/+archive/ubuntu/tools-proposed [10:32] ogra_: and that's pending promotion with next image promotion? [10:32] hmm, and even that is behind, i thinnk there is a 0.28 in wily already [10:32] * ogra_ checks -changes [10:32] ah, no [10:33] ogra_: right, cause we use that dtb now [10:33] lool, i'm not exactly sure how the promotion of udf works [10:33] i knwo it forst goes into that PPA before being released though [10:33] ogra_: I have a side question for you: how would you find livefs builds older than a week? the launchpad /+livefs page only lists less than ten or so [10:34] lool, there is an LP REST api for that (but i'm not sure it goes back further) and cjwatson could probably point you to an easy way to get more listed on the LP page [10:35] (read: i dont actually know) [10:35] ogra_: hmm I guess there might be links in the cdimage logs triggering the lp builds [10:35] there are, definitely [10:36] i would hope theer is an easier way though [10:39] it's likely LP API exposes it, I'm just lazy [10:40] (besides, I dont need it anymore if it's fixed in latest tools, it was just in case I'd have to dive into the issue :-) [10:40] rsalveti: confirmed fixed with tools-proposed, nevermind; marking fix-committed === clobrano is now known as c-lob [10:56] hmm I have to reboot to get network up [10:56] due to interface renaming it seems [10:57] hmm, i thought Chipaca had fixed that [10:57] (well, at least i know he was working on it ) === c-lob is now known as clobrano === clobrano is now known as c-lob [10:59] hmm I'm trying to enable SSH after booting an image without it; I see it's ssh_enabled: true in some yaml file, but can't figure out which [10:59] ah cloud-init userdata [11:04] ok, answer: sudo rm /etc/ssh/sshd_not_to_be_run and reboot (note for self) [11:07] GRRR [11:07] MMC: bcm2835_sdhci: 0 [11:07] reading uboot.env [11:07] *** Warning - bad CRC, using default environment [11:10] WHEE ! [11:10] MMC: bcm2835_sdhci: 0 [11:10] reading uboot.env [11:10] In: serial [11:10] that looks so much better :)) [11:13] grmbl ... [11:16] so switching to uboot.env on the rpi actually wipes the env of the preloader ... crap [11:21] EEEK ! [11:21] probably because nothing in the oem snap lands where it should [11:23] sergiusens, hulp ! [11:23] sergiusens, is there a way for me to have a whole subdir copied into the snap 1:1 [11:24] http://paste.ubuntu.com/11992407/ ... everything with "overlay" in the name should be in an overlay/ subdir [11:25] bah ... and it puts the totaly wrong snappy-system.txt in place, thats really bad [11:26] ogra_: hmm, there was no dir copy support, just files [11:27] we can add it, but it's not there [11:27] http://paste.ubuntu.com/11992517/ [11:27] hmm [11:28] well, the preloader has that subdir hardcoded, so if we ant to be able to use overlay dtb's it needs to exist [11:28] wasnt there a hack via the device tarball ? [11:28] ogra_: so all those files will be copied to the base, if you want a target, give it a target [11:28] * ogra_ thinks he remembers something like that + [11:28] in the -path: line ? [11:28] * ogra_ tries [11:31] ogra_: http://bazaar.launchpad.net/~phablet-team/goget-ubuntu-touch/trunk/view/head:/diskimage/common.go#L90 [11:31] Installing pi2_0.15_all.snap [11:31] stat /tmp/oem231909283/oem/pi2/0.15/boot-assets/overlays/ads7846-overlay.dtb overlays: no such file or directory [11:31] it should work, but 'should' is a strong word [11:31] ogra_: hmm, I can fix that [11:32] well, can i fix it from the snap side ? [11:32] ogra_: I can't think of anything [11:32] i kind of dont want to modify the stable released image :) [11:32] ogra_: oh, it's just one bug fix to u-d-f [11:33] not the actual image [11:33] ah [11:33] cool, yeah that will work [11:33] ogra_: to be able to deploy directories === utlemming is now known as utlemming_away === utlemming_away is now known as utlemming === utlemming is now known as utlemming_away [12:22] lool: both tools and tools-proposed are currently in sync [12:24] I sync the ppas when doing the stable release === utlemming_away is now known as utlemming [12:27] * ogra_ sighs ... [12:27] i'll try with the old rpi preloader after lunch ... thats depressing :/ === chihchun is now known as chihchun_afk [13:12] rsalveti: hmm I tried this like Friday with I believe up-to-date tools but perhaps I hadn't apt-get updated hmm [13:12] rsalveti: anyway, I've marked this fix committed [13:12] you can mark fix released if it's in wily too [13:12] (not sure what rule we follow there) [13:15] * ogra_ wonders why lool's uboot.env works but not mine ... life is so mean :P [13:22] ogra_: might be the time at which it's loaded? [13:23] ogra_: also, are you sure yours is loaded? [13:24] lool, yes, my prob is that i can get certain vars only from the preloader ... it hands them to u-boot once it loads it ... loading my env file wipes these vars, loadin yours keeps them [13:24] (well, loading yours using your uboot binary ... and loading mine using my uboot binary) [13:25] http://paste.ubuntu.com/11993082/ specifically "args" and "usbethaddr" are the ones i need [13:25] the only difference is that i bumped the uboot-env size to the same 128k that we use on the BB [13:26] lool, did you patch anything or did you just build using the defconfig ? [13:27] ogra_: just defconfig and the patch to allow raw initrd [13:28] oh, i forgot that one ! [13:28] but it's not needed in recent upstream versions [13:28] though i doubt that would unset atags or dtb values (or however the pre-loader hands them to uboot) [13:29] * ogra_ starts over and tries a completely unpatched build [13:33] sergiusens: do you see any problem with backporting the gocheck update? https://code.launchpad.net/~elopio/snappy/backport-gocheck/+merge/266663 [13:34] hmpf [13:34] U-Boot> printenv [13:34] fdtfile=bcm2836-rpi-2-b.dtb [13:34] usbethaddr=b8:27:eb:04:db:1c [13:34] still no args [13:36] jdstrand, what is the "mac-based isolation" security provided by snappy? [13:36] elopio: added to my todo [13:37] tx [13:41] ogra_: it's odd that you're not getting more in your environment [13:41] ogra_: it looks like you're not even getting the default environment [13:41] lool, i load an empty uboot.env [13:41] thats on purpose [13:41] ogra_: maybe the u-boot logic behaves differently in this case though in that you're not getting the default env like bootcmd etc. [13:41] (i first need to make sure the values from the preloader actually end up in my env) [13:41] ogra_: so perhaps you need to do a dump of the default environment and copy that in your env file [13:42] lool, yes thats what i'll do in the end [13:42] and thats how we worked on the BBB [13:42] but i need the intersection between the default env and the atags the rom hands ove [13:42] r [13:43] for the BBB i just loaded the emopty env file and got all the ROM values set [13:43] so i knew what to leave out of my env file (to have it "saveenv'ed" on first boot, since they are HW specific (MAC, seial etc) [13:44] i wanted to do the same on the rpi ... but seemingly these values are comletely gone now === chihchun_afk is now known as chihchun [14:39] fgimenez: lets skip today. [14:39] sorry. [14:47] elopio: ok np [15:33] rsalveti: I can land things in the 15.04 branch, right? [15:33] elopio: yes === chihchun is now known as chihchun_afk [16:10] any options as to mounting an NFS share on a snappy machine? [16:12] sarnold, jdstrand: what is the "mac-based isolation" security provided by snappy? [16:13] it puts all the ipads and iphones in a corner and tells them to be ashamed ? [16:14] :D [16:35] mterry: I assume this use of "mac" means mandatory access control [17:11] sarnold, not sure -- it's a phrase from https://developer.ubuntu.com/en/snappy/ [17:11] sarnold, I'm guessing so though [17:12] mterry: I guess the authors figured "mandatory access control" right next to "human friendly" didn't make sense. :) [17:13] sarnold, so it's basically apparmor looks like from a quick search? [17:14] mterry: sorry, I don't understa [17:14] understand [17:14] sarnold, I'm not familiar with "mandatory access control" either -- searching online for that phrase makes it sound like what apparmor does [17:15] mterry: ah! yes, apparmor is one implementation of a mandatory access control system, yes [17:15] yes, its a variant of mac [17:15] sarnold, jjohansen: cool, thanks! [17:15] specifically its a variant of dte instead of just te [17:15] a discretionary access control system allows data owners to set policies -- e.g., the unix permissions bits, acls, etc; mandatory access controls have a system adminstrator or security officer set permissions [17:19] * ogra_ definitely liked the naughty step explanation above more [19:04] mterry: it's true that tomcat wont start out of the box with example snap, but the mvn plugin itself works and I'd like to keep demoing it; would you mind merging the branch? [19:04] mterry: or I can remove the sample snap if that's comfort [19:04] (from the branch) [19:05] lool, OK, let me look again [19:10] mterry: the other way to fix it would be to launch tomcat ourselves (or fork catalina.sh) but I wanted to avoid this hardcoding and didn't have time to research why systemd wouldn't launch tomcat in the default script, sorry [19:10] mterry: I'm about to close laptop and leave for travel, worst case I'll demo from the branch and we can land it later [19:11] lool, I'm finding a few nits, yeah. Will comment. But if time is pressing, demo from the branch is fine [19:11] mterry: yeah, I have a fallback dont worry [19:11] lool, sorry didn't know this was a time-sensitive branch for ya [19:11] Or didn't realize when it would be at least [19:13] mterry: yeah, it's ok, I can do it from branch [19:14] mterry: reason I wnted to land it soon is because we're releasing a video to demo snapcraft with maven plugin, so would be nice if it was in tip [19:14] lool, commented [19:17] mterry: thanks [20:56] rsalveti: is there a pythononic way to know if you're on snappy? [20:57] rsalveti: python -c 'import platform; print(platform.dist())' doesn't distinguish between regular Ubuntu and Snappy Ubuntu