| === ezobn1 is now known as ezobn | ||
| === ezobn1 is now known as ezobn | ||
| fgimenez | good morning | 07:06 |
|---|---|---|
| dholbach | good morning | 07:10 |
| mvo | fgimenez, Chipaca, ricmm: image r152 should be good for testing | 07:58 |
| fgimenez | mvo, ok thanks, on it | 08:06 |
| fgimenez | mvo, 147 for armhf, right? | 08:08 |
| mvo | fgimenez: I have not build a armhf image with the latest snappy for 15.04 yet, I started that some minutes ago, wanted to verify on amd64 first | 08:13 |
| mvo | and there were a bunch of integration issues, but hopefully all good now | 08:13 |
| fgimenez | mvo, ok | 08:14 |
| mvo | fgimenez: there a some version (r151, r150) that are broken and will not work with system-image-cli correctly, just fyi when you do rollback testing | 08:14 |
| fgimenez | mvo, ok, should it work from r149? | 08:15 |
| mvo | fgimenez: yeah, that sounds good, from the latest you have faith in :) | 08:16 |
| mvo | fgimenez: to r152 or later and it should be good | 08:16 |
| fgimenez | mvo, :) ok i'll let you know how it goes | 08:17 |
| mvo | thanks | 08:17 |
| Chipaca | mo'in, whippersnappers! | 08:27 |
| * Chipaca adds a "whipper" subcommand | 08:27 | |
| mvo | hey Chipaca - new 15.04 is ready | 08:27 |
| Chipaca | mvo: so i hear :) congrats | 08:27 |
| Chipaca | mvo: i also hear it's been painful | 08:27 |
| mvo | thanks, was a hard fight | 08:27 |
| Chipaca | mvo: there is go generate in 1.4 | 08:30 |
| Chipaca | mvo: 15.04 has go 1.3 | 08:30 |
| Chipaca | oh dear | 08:31 |
| * Chipaca is pedantic today | 08:31 | |
| mvo | yeah | 08:31 |
| mvo | meh | 08:31 |
| JamesTait | Good morning all; happy Tuesday, and happy International Dot Day! 😃 | 08:32 |
| Chipaca | mvo: wait, it worked with xreadPassword and not with readPassword? | 08:33 |
| Chipaca | or am i misunderstanding somehting | 08:33 |
| * ogra_ notes that mvo did quite some imae builds :) | 08:33 | |
| mvo | Chipaca: yes | 08:36 |
| mvo | Chipaca: thats exactly it, I still have not tracked that down, but will soon | 08:36 |
| fgimenez | mvo, rolling back from 153 to 149 doesn't seem to work http://paste.ubuntu.com/12416211/ | 08:38 |
| mvo | fgimenez: hrm hrm hrm | 08:41 |
| * mvo tries to reproduce | 08:46 | |
| ogra_ | mvo, more kids ? oO | 08:47 |
| Chipaca | JamesTait: ⡧⢼⡮⢵⡯⡯⢱⠁ ⣏⠆⢎⡱⢹⠁⠮⡲⡃ | 08:47 |
| ogra_ | (and you really dont need to tell us) | 08:47 |
| JamesTait | Impressive, Chipaca! | 08:48 |
| Chipaca | JamesTait: sorry for not even trying to write "international" :) | 08:48 |
| Chipaca | ogra_: maybe we should tell him about getting off irc for that, also | 08:49 |
| ogra_ | Chipaca, true true ... germans ... tsk | 08:51 |
| Chipaca | ogra_: ooh! did you know "tsk" and "tut" are american and british ways of writing the same ("dental click") sound? | 08:52 |
| ogra_ | hah, nope :) | 08:52 |
| Chipaca | me neither | 08:52 |
| Chipaca | https://en.wikipedia.org/wiki/Dental_clicks | 08:53 |
| ogra_ | heh, that page makes me wanting to do a handstand to read it | 08:54 |
| ogra_ | all that phonetic upside down stuff :) | 08:54 |
| mvo | fgimenez: r152->153->152 seems to work, I'm downloading 149 now | 08:54 |
| ogra_ | looks like my amd64 installed system has properly updated from 148 to 151 via autopilot over night :) | 08:56 |
| fgimenez | mvo, ok i'll try that path too, the integration suite is passing on 153 | 08:59 |
| fgimenez | but including fake updates, i'll try it now with real updates and executing from a rollback | 09:00 |
| Chipaca | mvo: we don't need https://code.launchpad.net/~mvo/snappy/15.04-services/+merge/270965 now do we? | 09:06 |
| mvo | Chipaca: indeed, no | 09:07 |
| Chipaca | mvo: https://code.launchpad.net/~mvo/snappy/snappy-lp1490574-systemd-type-forking/+merge/270142 could use a bit of love | 09:08 |
| mvo | fgimenez: I think I found the issue, for some reason the boot-ok job seems to have not triggered | 09:11 |
| mvo | Chipaca: will do once I found the root cause for the rollback problem | 09:12 |
| Chipaca | ayup | 09:12 |
| mvo | pitti: did anything change between vivid and wily re "After=multi-user.taget"? I changed the ubuntu-snappy.boot-ok.service to have after=multi-user.target and that appears to be fine in wily (I can see that the service gets started and the output in wily but on vivid I don't see it starting, just that its active | 09:15 |
| pitti | mvo: "active" means that it started -- perhaps you mean "enabled"? | 09:16 |
| pitti | mvo: nothing wrt. After=, Requires= or enablement changed, but maybe the enablement link somehow got dropped in vivid? | 09:16 |
| mvo | pitti: http://paste.ubuntu.com/12416363/ is what I see on vivid | 09:17 |
| pitti | mvo: can you pastebin the journal? | 09:17 |
| fgimenez | mvo, great! thanks :) the rest of tests seem to be fine so far | 09:18 |
| pitti | mvo: that's not affected by the "too late bind mounting" issue, right? (as /etc/systemd/ gets bind-mounted in initramfs already0 | 09:18 |
| mvo | pitti: yeah, it gets bind mounted early | 09:18 |
| mvo | pitti: http://paste.ubuntu.com/12416373/ | 09:18 |
| pitti | mvo: systemctl cat ubuntu-snappy.boot-ok.service | 09:19 |
| pitti | mvo: find /etc/systemd -name '*boot-ok*' | 09:19 |
| pitti | mvo: nevermind | 09:20 |
| pitti | Breaking ordering cycle by deleting job ubuntu-snappy.boot-ok.service/start | 09:20 |
| mvo | pitti: ohhh | 09:20 |
| mvo | pitti: \o/ | 09:20 |
| pitti | mvo: so cat'ing the unit would help | 09:20 |
| ogra_ | pitti, just FYI https://launchpadlibrarian.net/217937303/initramfs-tools-ubuntu-core_0.7.7~ppa6_0.7.7~ppa7.diff.gz | 09:21 |
| mvo | pitti: http://paste.ubuntu.com/12416384/ <- its the wantedby=gettey, right? | 09:21 |
| mvo | hm, no | 09:21 |
| pitti | ogra_: nice! | 09:22 |
| ogra_ | pitti, i'm considering moving all the bind mounts to that next iteration | 09:22 |
| mvo | ogra_: I was wondering about this too, I don't think it matters much but why have two places if one is sufficient | 09:22 |
| ogra_ | right | 09:23 |
| pitti | mvo: ah, yes indeed -- getty.target wants boot-ok, but this runs after multi-user; maybe newer versions accept this kind of mixed wants/after cycle, but not yet vivid's? | 09:23 |
| pitti | mvo: or do you get a different cycle on wily, and it's just pure luck? | 09:23 |
| ogra_ | and who knows what other races we might cause by mounting half of them later | 09:23 |
| mvo | pitti: yeah, now I just need to figure out what Install synatx to use instead of getty.target | 09:24 |
| pitti | ogra_: do we have fsck of the /writable partition in initramfs? | 09:24 |
| pitti | mvo: multi-user.target doesn't work? | 09:24 |
| mvo | we do | 09:24 |
| mvo | pitti: simply wantedby=multi-user.target ? or removing the wantedBy entirely? | 09:24 |
| ogra_ | pitti, yeah, a few lines before the function above gets called | 09:24 |
| pitti | mvo: no, WantedBy=multi-user.target instead of getty.target (not sure why it's getty) | 09:24 |
| mvo | cool, will do | 09:25 |
| mvo | probably a historic accident | 09:25 |
| mvo | i.e. not talking to the right people :-D | 09:25 |
| * mvo hugs pitti | 09:25 | |
| ogra_ | pitti, http://paste.ubuntu.com/12416406/ ... handle_writable_paths does the mounting | 09:25 |
| ogra_ | the mount/umount to speed up e2fsck is a funny trick | 09:26 |
| ogra_ | (and works really well) | 09:26 |
| dholbach | tedg, do you know what's happening here? https://bugs.launchpad.net/snapcraft/+bug/1495525 | 09:27 |
| ubottu | Launchpad bug 1495525 in Snapcraft "qml plugin: example is broken" [Undecided,New] | 09:27 |
| mvo | Chipaca: I woder if the rollback issue could be releated to the new static grub.cfg | 09:35 |
| Chipaca | oh! | 09:35 |
| Chipaca | maybe? | 09:35 |
| * mvo looks deeper | 09:35 | |
| Chipaca | mvo: i'd suggest waiting for sergio on that one, as he'd know without looking :) | 09:36 |
| Chipaca | it's possible we said "no rollback from this change" | 09:36 |
| Chipaca | but i think the new static grub should deal with it as long as there's a versionless symlink | 09:36 |
| Chipaca | which there used to be, at leastr | 09:36 |
| mvo | Chipaca: hmm, looks like we have the old non-static grub.cfg in /boot/grub/grub.cfg :( | 09:38 |
| Chipaca | mvo: woo :-p | 09:42 |
| * guest42315 aw-food | 09:59 | |
| mvo | Chipaca: http://paste.ubuntu.com/12416556/ <- not looking great | 10:01 |
| ogra_ | mvo, uuh Bug 1495688 i see that too here (seems we have more fish to fry ) | 10:03 |
| ubottu | bug 1495688 in Snappy "system-image-cli tool fails on rolling/edge; missing file '/etc/system-image/config.d/00_default.ini" [Undecided,New] https://launchpad.net/bugs/1495688 | 10:03 |
| Chipaca | ogra_: there's a dangling symlink | 10:04 |
| ogra_ | hmm, not related to recent changes though ... my RPi image from sept. 2 has it too | 10:05 |
| ogra_ | Chipaca, yes, but why | 10:05 |
| ogra_ | it should ship client.ini | 10:05 |
| ogra_ | aha | 10:06 |
| ogra_ | should come from system-image-common | 10:06 |
| * ogra_ checks the manifest | 10:06 | |
| mvo | Chipaca: so we may need this services backport branch afterall | 10:06 |
| ogra_ | hmm | 10:07 |
| ogra_ | system-image-common is installed in the latest image | 10:07 |
| ogra_ | oh | 10:08 |
| Chipaca | mvo: D: | 10:08 |
| ogra_ | my phone only has system-image 2.5.1 ... snappy has 3.0.1 ... (both on vivid) | 10:08 |
| Chipaca | mvo: can we ship an override to update-grub that does something useful? | 10:10 |
| Chipaca | ohh | 10:15 |
| Chipaca | mvo: we're calling the _old_ update-grub on update? | 10:15 |
| Chipaca | or are we calling the new one? | 10:15 |
| mvo | Chipaca: we call the new one, maybe this could work http://paste.ubuntu.com/12416613/ | 10:16 |
| mvo | Chipaca: or simpler, we make the new update-grub write the static config and loose the ability to rollback (because the kernel will not be in /boot/grub/b | 10:17 |
| mvo | ) | 10:17 |
| dholbach | https://code.launchpad.net/~dholbach/snapcraft/fix-for-empty-stage_packages-in-plugins/+merge/271090 | 10:19 |
| dholbach | https://code.launchpad.net/~dholbach/snapcraft/small-typo/+merge/271086 | 10:19 |
| Chipaca | mvo: maybe we can make a release note about rollback not working | 10:20 |
| mvo | Chipaca: just making 09_snappy write the new static conf and stop would be ideal, least risky option but no rollback | 10:20 |
| ogra_ | mvo, oh ... do we need to seed system-image-snappy-common ? | 10:20 |
| mvo | ogra_: why? | 10:20 |
| ogra_ | mvo, system-image-common does not ship any client.ini file anymore | 10:20 |
| ogra_ | and i also assume we want the client.ini for snappy, not the phone one | 10:21 |
| mvo | ogra_: we can ship it in ubuntu-core-config then | 10:21 |
| mvo | ogra_: and we actually should do that already as 20_snappy iirc | 10:21 |
| ogra_ | fine with me too ... currently system-image-cli only throws exceptions | 10:21 |
| ogra_ | (on 15.04 edge that is) | 10:21 |
| mvo | ogra_: system-image-cli -v work for me on 154 (I also see the broken symlink) | 10:24 |
| ogra_ | mvo, and -i ? | 10:24 |
| mvo | ogra_: that explodes indeed | 10:25 |
| ogra_ | -v gets me a ton of permission denied errors (it shoudlnt) | 10:25 |
| ogra_ | hmm, ignore that ... happens on the phone too | 10:26 |
| ogra_ | i thought that uses the dbus backend so it doesnt need sudo | 10:26 |
| === kickinz1 is now known as kickinz1|lunch | ||
| fgimenez | mvo, should i file a bug about the rollback issue? | 10:46 |
| mvo | fgimenez: a good question, I think its understood, the hard questions is now what to do about it | 10:47 |
| sergiusens | dholbach, thanks for the branches! | 11:19 |
| sergiusens | dholbach, btw, have you read the new intro? I hope it is to your liking ;-) | 11:19 |
| ogra_ | workarounds for bug 1495688 and bug 1495683 uploaded | 11:24 |
| ubottu | bug 1495688 in Snappy "system-image-cli tool fails on rolling/edge; missing file '/etc/system-image/config.d/00_default.ini" [Undecided,New] https://launchpad.net/bugs/1495688 | 11:24 |
| ubottu | bug 1495683 in Snappy "Drop /etc/profile.d/Z99-cloud-local-test.sh from Cloud target" [Undecided,New] https://launchpad.net/bugs/1495683 | 11:24 |
| sergiusens | ogra_, they are called fixes ;-) | 11:32 |
| ogra_ | lol | 11:32 |
| dholbach | sergiusens, the google doc or updated docs in the branch? | 11:32 |
| sergiusens | dholbach, http://bazaar.launchpad.net/~snappy-dev/snapcraft/core/view/head:/docs/your-first-snap.md | 11:33 |
| dholbach | sergiusens, great - I'll check it out - thanks a bunch! | 11:33 |
| sergiusens | dholbach, I need to add a config walkthrough and extend that more (to get a bit more of a complicated example) ;-) | 11:33 |
| sergiusens | dholbach, but it has all the new stuff trunk has already | 11:33 |
| === kickinz1|lunch is now known as kickinz1 | ||
| dholbach | sergiusens, what's still on your list for the 0.2 release? | 11:40 |
| sergiusens | dholbach, I think we are good to switch already | 11:43 |
| sergiusens | dholbach, well, I want my config branch in and to extend the docs | 11:44 |
| dholbach | ok cool | 11:44 |
| ogra_ | mvo, triggering an image to make sure the file removals worked | 11:44 |
| dholbach | just asking to get an idea of what's planned :) | 11:44 |
| sergiusens | dholbach, soon after comes an interesting feature ;-) | 11:48 |
| dholbach | sergiusens, which one | 11:49 |
| dholbach | ? | 11:49 |
| Chipaca | anybody got a snap with an invalid yaml lying around? | 11:50 |
| Chipaca | wanting to test some error reporting | 11:50 |
| Chipaca | elopio: fgimenez: you, by chance? ^ | 11:51 |
| sergiusens | Chipaca, just unpack and repack ;-) | 11:52 |
| Chipaca | sergiusens: how do i repack? trying with "ar" i can never get it to work | 11:53 |
| fgimenez | Chipaca, there's an integration test for that, but iirc it generates the wrong yaml on the fly, let me check | 11:53 |
| Chipaca | and snappy build won't let me build a broken one :) | 11:53 |
| ogra_ | Chipaca, ar and snappy build indeed | 11:53 |
| ogra_ | oh | 11:54 |
| Chipaca | i'm wanting to build a snap with an *invalid yaml* :) | 11:54 |
| Chipaca | snappy build is not down with that | 11:54 |
| ogra_ | yeah, only read the last line after typing | 11:54 |
| Chipaca | i can build subtly broken ones, but not "aiee yaml parser kaput" broken ones | 11:54 |
| Chipaca | and that just happens to be an error case unlike others | 11:55 |
| sergiusens | Chipaca, I thought dpkg-deb would do the trick | 11:55 |
| Chipaca | in that it's a 500 error, from a yaml, in an async handler | 11:55 |
| Chipaca | trying dpkg-deb... | 11:56 |
| fgimenez | Chipaca, sorry nope, the test just check that actually you can't build it http://bazaar.launchpad.net/~snappy-dev/snappy/snappy/view/head:/_integration-tests/tests/build_test.go#L75 | 11:58 |
| Chipaca | fgimenez: sergiusens: ogra: dpkg-deb -R, and dpkg-deb --nocheck -b, do the trick | 11:58 |
| Chipaca | http://pastebin.ubuntu.com/12417202/ | 12:17 |
| Chipaca | \o/ | 12:17 |
| Chipaca | ^ that's what the rest api thinks of you having a fortune instead of yaml | 12:18 |
| Chipaca | (that's with my still in-progress "give context to errors" branch) | 12:19 |
| sergiusens | Chipaca, \o/ | 12:32 |
| ogra_ | hmm | 12:48 |
| ogra_ | neither /etc/system-image/config.d/00_default.ini nor /etc/profile.d/Z99-cloud-local-test.sh seem to exist at build time (i see the sccript to remove them being executed during build but it doesnt rm any files) | 12:49 |
| ogra_ | i wonder where they come from if not the build | 12:49 |
| * ogra_ twiddles thumbs waiting for the importer | 12:51 | |
| dholbach | davidcalle, which articles were you working on? | 13:18 |
| dholbach | davidcalle, is there any stuff which should go into the individual branches? or will it be just for developer.u.c? | 13:19 |
| Chipaca | darn, yesterday i was able to make the yaml parser panic, but today i can't reproduce it | 13:19 |
| Chipaca | should've written down the yaml | 13:19 |
| davidcalle | dholbach, not me directly, but some new snappy doc being written. Just duc :) | 13:19 |
| dholbach | ok | 13:19 |
| davidcalle | not *by me | 13:19 |
| tedg | dholbach: No, are you on vivid? Perhaps things changed in wily. | 13:21 |
| dholbach | tedg, I'm on wily, yes | 13:21 |
| dholbach | tedg, would it make sense to let qml snaps just depend on ubuntu-sdk-libs? | 13:22 |
| tedg | dholbach: No, not really. The SDK requires a lot more than we can support today. | 13:23 |
| dholbach | ok, I was just wondering it we could simplify the list of packages in snapcraft/plugins/qml.py somehow | 13:25 |
| rickspencer3 | does anyone know what I need to do figure out what to hw-assign so that my snap can use gpio pins on the rpi2? | 13:26 |
| rickspencer3 | jdstrand, ^ ? | 13:26 |
| rickspencer3 | apparmor only reports net_admin getting denied | 13:26 |
| ogra_ | sigh ! | 13:30 |
| mvo | sergiusens, Chipaca: do you think https://code.launchpad.net/~mvo/snappy/snappy-all-for-15.04 covers the bases? I want to tidy it a bit more and I think its good but a second opinion from an expert would be great | 13:30 |
| ogra_ | (amd64)ogra@aleph2:~$ ls -l /etc/system-image/config.d/00_default.ini | 13:31 |
| ogra_ | lrwxrwxrwx 1 root root 13 Sep 15 15:26 /etc/system-image/config.d/00_default.ini -> ../client.ini | 13:31 |
| ogra_ | (amd64)ogra@aleph2:~$ | 13:31 |
| * ogra_ doesnt get where it comes from, it doesnt seem to exist at rootfs build time | 13:31 | |
| sergiusens | mvo, you don't like asking questions with simple answers, eh? :-) | 13:32 |
| Chipaca | mvo: i agree, we need an opinion from an expert. | 13:33 |
| Chipaca | mvo: we're going to have to go find one | 13:33 |
| mvo | lol | 13:33 |
| ogra_ | do we have spare headcount to hire one ? | 13:33 |
| ogra_ | job description "expert" | 13:34 |
| mvo | I think its good, it just needs its own (blocking) systemd-job to ensure everything is in place before the user can hit the commandline | 13:34 |
| === ezobn1 is now known as ezobn | ||
| ogra_ | GRRR ... so calling [ -e ... ] on a dangling symlink isnt the cleverest of things | 13:34 |
| tedg | dholbach: I think it does make sense to do some sort of meta-package, but the SDK has too much other stuff in it. | 13:35 |
| * ogra_ switches to -L | 13:35 | |
| mvo | ogra_: if you have spare cycles, system-image.ubuntu.com//pool/device-f6dbcd89372ad1bd2b6a37df25b2e02fae38530d92c9ebd541bf327c27048794.tar.xz has also some danling symlinks in assets/vmlinuz | 13:35 |
| * mvo dives back into grub | 13:35 | |
| dholbach | tedg, I thought ubuntu-sdk-libs was just the stuff phone apps could know to expect around | 13:36 |
| ogra_ | ARGH ! | 13:36 |
| Chipaca | mvo: that code seems correct, but i'd test it all the ways | 13:36 |
| ogra_ | copy pasting a filename with typo out of a bug report isnt so clever either | 13:36 |
| * ogra_ should take the day off | 13:37 | |
| tedg | dholbach: Sure, but we don't want all of those things for a QML snap. We can't provide all the services that they require. That's gotta be provided by the U8 framework. | 13:37 |
| dholbach | ok | 13:38 |
| rickspencer3 | jdstrand, could you take a look and let me know why I am seeing the denial here, considering what is in my apparmor profile? | 13:41 |
| rickspencer3 | http://pastebin.ubuntu.com/12417599/ | 13:41 |
| jdstrand | rickspencer3: sorry, in a meeting | 13:41 |
| jdstrand | be bak in a minute | 13:41 |
| rickspencer3 | k | 13:41 |
| jdstrand | ok, back | 13:49 |
| jdstrand | rickspencer3: instead of /sys/devices/platform/soc/3f200000.gpio, use /sys/devices/platform/soc/3f200000.gpio/** | 13:51 |
| ogra_ | ok ... since all good things in life are three ... | 13:52 |
| * ogra_ triggers the third image for that file removal stuff | 13:52 | |
| rickspencer3 | ok, I'll try that | 13:54 |
| rickspencer3 | jdstrand, the only difference is the *? | 13:58 |
| rickspencer3 | because I am still seeing the denial | 13:58 |
| rickspencer3 | jdstrand, I just added /* to the app armor profile and reparsed it | 13:58 |
| rickspencer3 | I didn't do hw-assign again | 13:58 |
| ogra_ | it is /** not /* | 13:59 |
| jdstrand | rickspencer3: use '**'. not '*' | 13:59 |
| jdstrand | s/\.// | 13:59 |
| rickspencer3 | oops, text wrapping in xchat :/ | 14:00 |
| jdstrand | but yes | 14:00 |
| jdstrand | the /sys/devices/platform/soc/3f200000.gpio/** means 'everything under (but not including) /sys/devices/platform/soc/3f200000.gpio' | 14:01 |
| mvo | pitti: I assume "before=multi-user.target" and "wnatedby=multi-user.target" is safe and will ensure my service has finished before a user can login? | 14:01 |
| fgimenez | ogra_, with these changes https://code.launchpad.net/~fgimenez/snappy/resize_test2/+merge/271131 the resize test works (and passes) on 15.04, r155 :) | 14:22 |
| ogra_ | fgimenez, yay | 14:22 |
| mvo | ogra_: http://paste.ubuntu.com/12417786/ should hopefully fix the broken symlinks in assets/ that I saw earlier. its basicly just a forward port of the wily code, but a second pair of eyes would be great | 14:23 |
| ogra_ | mvo, can we copy the config too if we copy abi and System.map ? | 14:24 |
| mvo | yeah | 14:25 |
| ogra_ | mvo, looks fine otherwise | 14:25 |
| ogra_ | i still fight with /etc/system-image/config.d/00_default.ini :( | 14:26 |
| ogra_ | i dont get why | 14:28 |
| pindonga | mvo, hi there, remember that bug about supporting forking services in systemd unit files? I saw the mp pending review... is there anything blocking that? | 14:33 |
| mvo | just lack of tme, shoud land rsn | 14:38 |
| pitti | mvo: before/wantedby is the usual combination indeed | 14:38 |
| pitti | mvo: but stuff in multi-user.target can happen after gettys; if you strictly want "before users can log in", you want Before=systemd-user-sessions.service (WantedBy=multi-user.target is still okay) | 14:39 |
| pitti | mvo: the thing that seems to cause loops/trouble is if you order (After=) a service *past* a target it's wanted in; before is always fine | 14:40 |
| mvo | thanks | 14:40 |
| clobrano | hi there, I'm looking for info sources about installing udev rules in snappy (other than using hw-assing). Any suggestion? :) | 14:41 |
| pindonga | mvo, ack, thx | 14:49 |
| sergiusens | Chipaca, mvo dholbach https://code.launchpad.net/~sergiusens/snapcraft/wiki/+merge/271140 | 14:54 |
| fgimenez | ogra_, on bbb i'm getting this while executing the resize test http://paste.ubuntu.com/12417958/ | 14:55 |
| fgimenez | ogra_, this is the script used http://bazaar.launchpad.net/~fgimenez/snappy/resize_test2/view/head:/_integration-tests/scripts/install-test-initramfs | 14:57 |
| ogra_ | fgimenez, thats not the initrd | 14:58 |
| ogra_ | you want the one under /boot/uboot | 14:58 |
| fgimenez | ogra_, ok i'll fix that, thanks! :) | 14:59 |
| dholbach | sergiusens, replied | 15:03 |
| sergiusens | dholbach, thanks, replied (I wasn't part of the discussion either fwiw) :-) | 15:07 |
| dholbach | :) | 15:07 |
| dholbach | sergiusens, the summary of the example package is still a bit weird, but apart from that it looks good to me | 15:13 |
| sergiusens | dholbach, oh, if you have a better suggestion; just go ahead :-) | 15:16 |
| dholbach | maybe I just don't understand it :) | 15:16 |
| dholbach | sergiusens, is it meant to say "downloader"? | 15:16 |
| dholbach | or "downloaded"? | 15:17 |
| dholbach | I mean I'm nitpicking - the rest looks fine AFAICT :) | 15:17 |
| sergiusens | dholbach, yeah, the package name is downloader and it downloads from the wiki to create it | 15:17 |
| sergiusens | dholbach, I should of not worked on a downloader snap for this I guess :-) | 15:18 |
| sergiusens | dholbach, oh, but now I see the typo | 15:18 |
| dholbach | <3 | 15:18 |
| sergiusens | :-) | 15:18 |
| sergiusens | dholbach, now that's fixed | 15:20 |
| dholbach | sergiusens, approved | 15:21 |
| fgimenez | Chipaca, is this a good source of info for integration tests of the api http://bazaar.launchpad.net/~snappy-dev/snappy/snappy/view/head:/daemon/api_test.go ? is there anything else not covered there? | 15:41 |
| mvo | pitti: hi, sorry for bothering you again, I added a new service to snappy, its configured exactly like the boot-ok service and yet http://paste.ubuntu.com/12418278/ | 15:49 |
| pitti | mvo: any cycle in the journal? can you please cat the unit? | 15:52 |
| mvo | ogra_: more ininitramfs troubel it seems, /etc/systemd/system is a writable path, it contains a symlink in getty.target.wants to boot-ok which I removed from the image but because its a writable path with transition the file is never removed | 15:52 |
| pitti | mvo: ah, but this one is disabled too -- did you actually systemctl enable (or update-rc. enable) it? | 15:52 |
| mvo | pitti: http://paste.ubuntu.com/12418301/ is the journal | 15:53 |
| ogra_ | mvo, yeah, what i said during the meeting :) | 15:53 |
| ogra_ | we dont have any way to remove such files yet | 15:53 |
| mvo | ogra_: I bet thats the same issue for migrate-grub, its a new file in multi-user.tar.get.want but its not copied up too | 15:54 |
| pitti | mvo: ah, so I suppose that's related to "can't remove symlinks"? | 15:54 |
| mvo | pitti: http://paste.ubuntu.com/12418313/ | 15:54 |
| pitti | mvo: I suppose you just didn't enable this then? | 15:55 |
| mvo | pitti: yeah, I think its exactly this, /etc/systemd/system/multi-user.target.wants is in mode "transitional" but after the initial copy I think it never updates new files or deletes files so we are lost | 15:55 |
| mvo | pitti: it should be enabled via the deb, let me double check | 15:55 |
| ogra_ | \o/ ... | 15:56 |
| ogra_ | ok, the most recent rootfs only has 20_snappy.ini in the system-image/conf.d dir | 15:56 |
| mvo | pitti, ogra_: so the image has the new ubuntu-snappy.grub-migrate.service but the bind mount over it hides it. thats pretty terrible if we can not add or change job and that we can't fix bugs like that there is a wrong getty.target.wants | 15:58 |
| * mvo scratches head | 15:58 | |
| ogra_ | hmm, did you try setting the dir to sync instead of transition ? | 15:59 |
| ogra_ | (and reboot) | 15:59 |
| mvo | ogra_: I cross all fingers that I have | 16:00 |
| ogra_ | well, else we need a hack in initrd | 16:00 |
| ogra_ | actually it might be cool if we could use rsync instead of cp in that code ... that might fix some issues | 16:01 |
| mvo | ogra_: no, synced does not help at all :/ | 16:05 |
| ogra_ | :( | 16:05 |
| seb128 | mvo, you noticed that the snappy update in wily is depwait on goget-ubuntu-touch which depwait on golang-uboot-go-dev? | 16:06 |
| mvo | ogra_: hm, looking at the source synced should work, it should at least make my files appear | 16:10 |
| ogra_ | yes | 16:10 |
| ogra_ | mvo, even transition should work | 16:11 |
| * ogra_ doesnt get it | 16:13 | |
| ogra_ | i wonder if we could force it with an extra "cp -u $redaonly-dir $writable-dir" | 16:14 |
| mvo | ogra_: so /etc/systemd/system/getty.target.wants/ubuntu-snappy.boot-ok.service needs to go and /etc/systemd/system/multi-user.target.wants/ubuntu-snappy.{grub-migrate,boot-ok}.service must appear | 16:20 |
| === kickinz1 is now known as kickinz1|eod | ||
| === kickinz1|eod is now known as kickinz1 | ||
| ogra_ | mvo, yeah, not sure how we should do that ... it seems to never update anything actually | 16:22 |
| ogra_ | mvo, if i unmount the dir i also see a webdm service that got never copied | 16:23 |
| mvo | ogra_: yeah, I'm not suggesting to hack it, jsut do document the issue | 16:24 |
| mvo | ogra_: pretty alarming | 16:24 |
| ogra_ | yes | 16:24 |
| cyphermox | ogra_: ricmm: could you please let me know when we have a rolling image with fwupdate included? | 21:07 |
| cyphermox | we'd shoud probably include fwupd too. | 21:07 |
| cyphermox | zomg, I can't type. | 21:07 |
| ogra_ | whats that ? | 21:08 |
| ogra_ | (firmware ? firewall ? ) | 21:08 |
| ogra_ | (funny witches) | 21:09 |
| cyphermox | firmware. | 21:11 |
| cyphermox | disregard fwupd for now since it's not even in Debian. | 21:11 |
| ogra_ | heh | 21:17 |
| ogra_ | cyphermox, thats for bluez ? | 21:17 |
| cyphermox | no. it's for firmware updating magic in the EFI specification | 21:18 |
| cyphermox | as in, permanent updates :) | 21:18 |
| ogra_ | ah, shudder ,... you said EFI | 21:18 |
| cyphermox | it's the future. | 21:18 |
| * ogra_ pets uboot | 21:18 | |
| * cyphermox kicks uboot :) | 21:19 | |
| ogra_ | hey ! | 21:19 |
| ogra_ | :) | 21:19 |
| ogra_ | our grub setup is closer to uboot than to anything else nowadays :) | 21:19 |
| sergiusens | \o/ | 21:20 |
| ogra_ | \o/ | 21:20 |
| * ogra_ joins the yoga class | 21:20 | |
| cyphermox | ogra_: you can do some very cool magic with grub and EFI. | 21:20 |
| ogra_ | nah, i cant :) | 21:20 |
| ogra_ | but i can do some very cool magic with uboot :) | 21:21 |
| cyphermox | ok then, I can :) | 21:22 |
| ogra_ | heh | 21:22 |
| ogra_ | yeah, and i'm happy i know whom to poke if it comes to grub ;) | 21:22 |
| ogra_ | tough in regard to snappy ... | 21:23 |
| ogra_ | (amd64)ubuntu@localhost:~$ wc -l /boot/grub/grub.cfg | 21:23 |
| ogra_ | 54 /boot/grub/grub.cfg | 21:23 |
| ogra_ | we dont really have much there :) | 21:24 |
| ogra_ | (static file btw) | 21:24 |
| cyphermox | feel free to ask me | 21:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!