=== jamesh_ is now known as jamesh [07:12] morning [07:46] PR snapd#9921 closed: boot: helper for setting up a try recover system [07:49] good morning [08:04] morning [08:05] pstolowski: mvo: hey [08:08] good morning pstolowski and mborzecki ! happy monday [08:11] o/ [08:18] good morning mvo :) [08:29] good morning zyga, happy monday to you as well! [08:36] going downstairs to set up some recording, I'll be back in the afternoon [09:18] oh well, snapshot tests mock too much... [09:57] anyone seen this? https://paste.ubuntu.com/p/Vb9jzgR5Zx/ [10:16] mvo: ^^ looks like core-initrd was changed and our kernel repacking is broken now [10:16] mborzecki: it was changed, yes [10:17] mborzecki: check https://github.com/snapcore/core-initrd/commits/main [10:17] that's probably the changes for having more modules in the server builds [10:17] mvo: yeah, i suspect it's this commit: https://github.com/snapcore/core-initrd/commit/9ca7436f821f0e2d37874fd217407275f1e24e3c [10:18] mborzecki: yeah [10:37] PR snapd#10008 opened: tests/lib/prepare: fix repacking of the UC20 kernel snap for with ubuntu-core-initramfs 40 <âš  Critical> [10:39] mvo: ^^ [10:45] mborzecki: thanks, checking [10:46] mborzecki: I cced you on a question about appamror confinement partil vs full, I think I got the right answer but if it looks strange let me know (no need to deep dive) [10:47] mborzecki: \o/ thanks for the fix! [10:59] mvo: it looks correct [11:13] PR snapcraft#3464 closed: meta: add support for `kernel.yaml` for kernel snaps [11:37] PR snapd#9994 closed: tests: improve tests documentation - part 2 [11:49] mvo: pushed a couple more fixes to #10008 [11:49] Bug #10008: Automatic configuration of wireless pcmcia card does not work [11:49] PR #10008: tests/lib/prepare: fix repacking of the UC20 kernel snap for with ubuntu-core-initramfs 40 <âš  Critical> [12:13] PR snapcraft#3465 opened: [backport] store: improve platform detection (#2931) [12:19] mvo: #10007 is ready for reviews [12:19] Bug #10007: HD drive very slow when no cd inserted whith kernel 2.6 and Gnome 2.6 or 2.8 [12:19] PR #10007: o/configstate, o/snapshotstate: fix handling of nil snap config on snapshot restore [12:35] morning folks [12:36] o/ [12:36] hey diddledan [12:37] the snap store is being a pain (has been for days) https://usercontent.irccloud-cdn.com/file/Jf25WrzK/image.png [12:38] yeah entirely possible, we had to push out a new snapd/core snap which has a tendency to slow down things, hopefully this week it is better though [12:38] I'm sure my snap being 2.5 GB isn't helping [12:39] (CUDA libraries are huge!) [12:39] oh yeah that definitely is a problem, I seem to recall that snaps over 1 GB are very finnicky to upload [12:46] tis a good job I don't have a data cap with the number of times I've got to 99% [12:47] haha yeah that would really sting if you had a data upload cap too [12:47] mvo: not sure when I mentioned it last week, but for example today I see we only have one single PR running via github actions workers: https://github.com/snapcore/snapd/actions?query=is%3Ain_progress but many, many queued jobs [12:47] I think we have some github actions self hosted runners offline [12:49] self-hosted runners for the win though (I have one here at home ;-)) [12:49] yeah they are great except when you realize you now are doing devops :-| [12:50] true [12:50] [12:51] haha [13:13] ijohnson: yeah, not sure how to even see how many runners are up at this point [13:13] mvo: can you see it? [13:13] diddledan, you should try to find that norwegian guy (peer) that resets your connection and tell him to stop that ! [13:13] mborzecki: yeah real annoying with that u-c-i bug you found this morning too [13:13] ijohnson: and it's stuck in the actions queue ;) [13:13] peer is an ass [13:13] totally ! [13:18] ijohnson, hi [13:19] ijohnson, did you see this error= https://github.com/snapcore/snapd/pull/9997/checks?check_run_id=2055837552#step:5:547 [13:19] PR #9997: tests: update details on task.yaml files [13:19] hey cachio_ let me have a look [13:20] it started on friday I think [13:20] cachio_: that is the issue that mborzecki is fixing, but we are stuck waiting for github actions workers to run this PR: https://github.com/snapcore/snapd/pull/10008 [13:20] PR #10008: tests/lib/prepare: fix repacking of the UC20 kernel snap for with ubuntu-core-initramfs 40 <âš  Critical> [13:20] I already fixed the github actions agents [13:20] mvo, hi, is any of the agents offline? [13:21] ijohnson, ah, nice [13:22] I'll check again the workers [13:22] I fixed that on friday [13:23] thanks cachio [13:32] cachio_: I see a lot of prodstack offline, eg prodstack-a-45 to 76 (some isolted idle ones in their though) [13:33] mvo, thnaks [13:33] qc [13:34] ijohnson, mborzecki I can see some stat about the runners but not super many, I see a good bunch of stuff online but also quite a few offline [13:34] mmm that's unfortunate [13:36] hey guys [13:37] oh, workers down? [13:37] something like that [13:37] I learned one interesting thing the other day [13:38] gitlab can run CI for projects that are hosted on githu [13:38] *github [13:38] so you can use the somewhat more powerful pipeline system and easily host your workers (and scale them) [13:38] as the provisioning process is a bit better there [13:38] you could even auto-scale with juju [13:55] yeah maybe gitlab is nice, but I doubt we have any time to switch [13:56] a downside of gitlab-ci is there is no inbuilt concept of a `matrix` [14:00] diddledan that's true but I don't see that as a major flaw or limitation [14:00] you can easily create that with .extends [14:00] I think that having a native matrix would be nice to unclutter more complex pipelines [14:01] but I'm still learning that, and there are parts I did not explore [14:01] like pipeline triggering and subscription, that allow for even more composition [14:02] I like the ability to include/import pipeline stuff from external files/repos [14:02] that really helps when you have multiple projects that all need the same or similar pipelines [14:04] yeah, that's true [14:04] salsa has a very nice and featureful pipeline for packaging [14:04] in my current work I'm setting up reusable pipelines that are easy to include into various repos [14:05] it's not far off from github actions [14:05] but actions are not so composable as you can only define one step, and not a larger chunk === cachio_ is now known as cachio [16:10] * cachio lunch [17:55] ijohnson, hey, about the PR #9960 do you have any update? [17:55] PR #9960: tests: update permission denied message for test-snapd-event on ubuntu 2104 <â›” Blocked> [17:55] should we skip that test? [17:55] cachio: no not yet, but we should not skip that test on 21.04 [17:55] cachio: is that test blocking beta certification of 2.49.1 ? [17:56] ijohnson, no [17:56] yeah let's leave the test as-is for now, I will check in on it today after lunch [17:56] as 21.04 is in red [17:56] I want to know what to do with that one [17:56] ijohnson, I'll wait in that case, thanks [18:39] popey: if I register cuda-runtime-11-2-1804, would it be ok to organise to add it to snapcrafters with a view to making it a content-snap for anybody to use? [18:40] (I'm trying to work around the store bailing out because my snap is huge) [18:42] well I registered it anywho :-p [19:49] ijohnson, hi [19:49] quick question [19:49] hey cachio [19:49] I install the snap test-snapd-sh and then I remove it [19:50] it is ok if: [19:50] ls /sys/fs/cgroup/devices/snap.test-snapd-sh.sh/ [19:50] cgroup.clone_children cgroup.procs devices.allow devices.deny devices.list notify_on_release tasks [19:50] it remains there? [19:51] cachio: that's a bug where we don't remove/cleanup the device cgroup for a snap after it's removed [19:51] I filed that a while ago, let me see if I can find the bug [19:52] cachio: see https://bugs.launchpad.net/snapd/+bug/1803210 [19:52] Bug #1803210: snap's device cgroup is not discarded upon uninstall [19:52] ijohnson, ok, because I have a test failing and not sure if it could be related to that [19:53] cachio: how is the test failing? [19:53] ijohnson, https://paste.ubuntu.com/p/nWQS6DnZFd/ [19:54] don't know if the test clean up is bad [19:54] cachio: yeah we probably need that test to clean itself up [19:54] or something is not working properly when we remove a snap [21:58] PR snapcraft#3466 opened: store: do not unnecessarily catch/rewrite exceptions [22:55] PR snapd#10009 opened: boot: export bootAssetsMap as AssetsMap [22:55] PR snapd#10010 opened: boot/assets.go: add InstallObserverOptions to allow providing previous assets [23:58] PR snapcraft#3467 opened: gitignore: sort list [23:59] well fooey