[06:30] morning === pstolowski|afk is now known as pstolowski [08:05] morning, hey mborzecki [08:06] pstolowski: hey [08:18] mborzecki: should we disable opensuse tests for now? [08:18] pstolowski: still failing? [08:19] mborzecki: checkingm restarted one of the failing PRs [08:19] *checking [08:23] pstolowski: looks like it was timeouting a bit too [08:23] anyways, restarted #7702 and if it fails again on tw, we can switch it to manual [08:23] PR #7702: tests: adding fedora 31 [08:36] https://forum.snapcraft.io/t/cant-install-snaps-in-clean-16-04-lxd-container/14140 [08:36] this seems serious [08:37] probably due to snapd being super old in 16.04 and not supporting something new in latest core? [08:38] ExecMount=/bin/mount /var/lib/snapd/snaps/core_8039.snap /snap/core/8039 -t squashfs -o nodev,ro (code=exited, status=32) [08:38] the error is pretty generic, according to manpage it's '32 mount failure' [08:38] it's very easy to reproduce [08:39] if mount cannot mount then we're in deep trouble [08:40] popey: what was the host os? [08:40] IIRC there couldnt be a hard dependency on squashfuse in 16.04 [08:40] you need to install it manually [08:40] no ogra [08:41] this worked up until recently [08:41] it's a cron job I run every day [08:41] and today it failed [08:41] oh, ok [08:41] mborzecki: 19.10 [08:42] popey: was lxd installed from snaps too? [08:42] on the host, yes, stable [08:44] popey: ok, trying it out now [08:48] can you try the mount command manually ? that pwerhaps gets you more output [08:51] heh, can't start the container for some reason, and no logfile either [08:52] https://paste.ubuntu.com/p/nNmgJf6Ndf/ [08:52] works for me [08:53] (on a xenial host ... but not on recent kernel (havent rebooted for a while) [08:53] ) [08:56] lxc launch ubuntu:16.04 complains that no image was found, but images:ubuntu/xenial seems to work [09:10] PR snapd#7734 opened: spread: do not explain errors [09:11] pstolowski: super trivial ^ [09:29] PR snapd#7735 opened: spread: move openSUSE Tumbleweed to unstable systems [10:11] PR snapd#7736 opened: tests/main/system-usernames: Amazon Linux 2 comes with libseccomp 2.4.1 now [10:16] so in my vm, once i launched the container, snapd doesn't even start [10:20] pstolowski: hm looks like we'll need to merge 7736 and 7735 into one PR [10:21] mborzecki: hah, right [10:22] PR snapd#7736 closed: tests/main/system-usernames: Amazon Linux 2 comes with libseccomp 2.4.1 now [10:29] popey: btw. eoan is on 5.3 kernel? not sure why it's showing 5.0 in snap version in the output you pasted [10:29] and once i resized the vm image things work well now [10:59] mborzecki: that was inside the container [11:00] oh, you're right [11:00] my system has been upgraded from 18.04, maybe I am missing a package [11:03] something may have gone screwy in my upgrade, installing linux-image-generic now pulls in 5.3, gonna reboot and test again [11:29] that didn't help [11:30] mborzecki, hey [11:30] cachio: hey [11:30] left a coment in the PR #7735 [11:30] PR #7735: spread, tests: openSUSE Tumbleweed to unstable systems, update system-usernames on Amazon Linux 2 [11:30] yesterday I updated the image [11:31] but I think it is the same [11:31] cachio: heh, that's unfortunate [11:33] popey: is the 16.04 lxd image something you created in the past and keep updating? the one i pulled today had snapd 2.40 in it, rather than 2.27 [11:33] i guess. I dunno how to update it [11:34] cachio: the problem with system-usernames and opensuse 15.1 is differnt, looks like there's no daemon user at all [11:36] popey: so snapfuse is not in the contained, but do you have squashfuse there at least? [11:36] no [11:37] i have removed lxd and am re-installing to see if i get a new image [11:37] mborzecki, it could be because yesterday I imported the new opensuse leap 15.1 image [11:37] from opensuse project [11:38] mborzecki, I can revert that image if it is a problem [11:38] cachio: is the daemon:daemon user/group added manually when you prepare the image? [11:39] mborzecki, nottrobin [11:39] no [11:39] popey: still, makes me wonder how it worked before, since loopback isn't avaialbe inside the container and none of the fuse squashfs handlers are avaialble [11:44] mborzecki, I have a debug session opened [11:45] mborzecki, do you want to take a look? [11:45] cachio: https://doc.opensuse.org/documentation/leap/startup/single-html/book.opensuse.startup/index.html#sec-yast-userman-defaultsystemusers lists bin, daemon as some default users added for legacy apps [11:45] cachio: but i see neither in /etc/passwd [11:47] mborzecki, I am checking that in the base image that I got from opensuse project [11:50] mborzecki, in the mew image from opensuse cloud it is not included neither [11:55] cachio: w8, so we had 15.1 image before, but the test broke only after the last update, is that what happened? [11:55] mborzecki, yes [11:56] I iincludede the new image provided by opensuse cloud [11:56] cachio: then we should revert to the last good one and report a bug to whoever publishes the build [11:56] s [11:56] mborzecki, sure [11:58] mborzecki, reverted [11:58] and restarted tests [11:59] mborzecki, that should fix the issue [11:59] cachio: thanks [12:04] cachio: downloaded the latest 15.1 kvm image from opensuse.org and there's no bin or daemon users either [12:05] mborzecki, well, perhaps we should update our tests in that case [12:05] cachio: i mean, if it's the case, then the system-users feature isn't universal anymore and probably needs extra tweaks to be documented [12:06] mborzecki, agree [12:06] I'll research a bit more about this [12:41] Noob question, how do I actually add a group to my user? usermod -a -G isn't working since read-only system [12:41] pokk, you dont ... [12:41] (assuming you speak about ubuntu core) [12:41] what do you try to do ? [12:42] (why do you think you need the group= [12:42] ) [12:44] hi [12:46] is it about snap? Ubuntu new toy [12:46] this channel is for snap development, snapd development and ubuntu-core [12:47] PR snapd#7735 closed: spread, tests: openSUSE Tumbleweed to unstable systems, update system-usernames on Amazon Linux 2 [12:47] (also ... not actually new, snap packages exist since 2014) [12:48] what is this channel about? [12:48] (nor ubuntu ... snaps can be installed on a ton of different distros) [12:48] ok snap development ok, i thought you are a bot [12:49] heh ... said the person that joins and leaves the channel between two lines [12:52] looks like jean claude van damme is busy on freenode again ... [12:53] ... doing net-splits ... [13:17] yawn [13:17] good morning [13:18] how are things? [13:22] how's the sleep situation ? [13:25] ogra: I totally fall asleep around 9PM [13:25] ogra: and then wake up 3-4AM [13:25] so... okay I guess [13:34] bug triage duty is great when LP keeps timing out ;) === ricab is now known as ricab|lunch [13:48] hey zyga! [13:48] hey :) [13:48] how are you doing? [13:48] mvo can't sleep either i suppose, is active in the standup doc ;) [13:48] it's 5:48, I wish I could sleep normally [13:48] haha [13:49] yeah [13:49] it's the same for everyone here [13:49] zyga: it's going fine here. i'm warming up for the new macbook.... ;) [13:50] zyga: but not this year [13:50] yeah [13:50] I will first try to sell my 2018 [13:50] if that succeeds I'll go for it [13:50] if not I'll stick around for longer [13:50] zyga: just realized it gets much more attractive with VAT return :) [13:50] pstolowski: yeah [13:50] pstolowski: if you go to an apple reseller they almost always get you better price too [13:51] pstolowski: just start talking to them [13:51] i keep forgetting about that.. it's significant chunk in this case in particular [13:51] pstolowski: it's not the stock price in the end [13:51] pstolowski: vat, income tax, extra cut if you negotiate for a sec [13:51] pstolowski: and that adds up [13:51] pstolowski: I really just don't want two laptops at once :) [13:52] pstolowski: I was thinking and the 6 core, 1TB, 32GB model is what I would look at most likely [13:52] zyga: yeah that's my problem too... too much hardware already at home [13:52] pstolowski: probably with the GPU upgrade since it's cheap but significant [13:53] pstolowski: I want to play some music on it because apparently the speaker upgrade is incredible (starting from a already fantastic quality of 15") [13:53] pstolowski: there are 6 speakers in that thing now [13:53] I was out for a walk last night but went to the seaside instead [13:53] zyga: yeah, 1TB + 32GB would be my setup too. i don't care about GPU too much [13:53] pstolowski: the GPU is just $100 but it's not something you can ever upgrade [13:54] zyga: yes, i watched a review, speakers are super good apparently [13:54] I'd get it because it's useful to have vram for VMs as well as apps on the host [13:54] I don't remember but I think the extra $100 gets you 8GB of vram [13:55] zyga: ah, if you get more VRAM with that.. i though VRAM was fixed, just two chipsets to choose from [13:55] I could be wrong, let's check [13:55] then yes, by all means [13:55] oh, there are 3 GPU tiers [13:55] that's ... weird [13:55] I checked on the website in the US [13:55] and I saw two [13:56] but on my phone with the apple store app I see three (tied to my eu account) [13:56] 4GB 5300M, 4GB 5500M and 8GB 5500M [13:56] anyway [13:56] that's all theoretical [13:56] zyga: you'll still end up carrying x240 :) [13:57] mborzecki: yeah but that's different [13:57] mborzecki: I would replace my 15" with the 16" because at home it's just awesome to use [13:57] mborzecki: it's really my most loved device over the years [13:57] mborzecki: despite the issues actually [13:57] mborzecki: x240 is something I can comfortably throw on the xray scanner at the airport [13:57] idk, i have no feelings for devices i own ;) [13:57] haha [13:57] mborzecki: and use for one week without missing things I like normally [13:58] mborzecki: you say that because you are missing out ;-) [13:58] haha [13:58] ;D [13:58] that! [13:58] * zyga is totally silly now [13:58] surprised they don't treat thinkpads as possible weapons [13:58] pstolowski: did you notice the new github app for ios? [13:58] it's invite only for now but it is coming [13:58] zyga: nope, for ios? [13:58] yep [13:58] android later as well [13:59] hmm what is it good for? [13:59] everything normal github is good for but tailored for smaller screen [13:59] code reviews? [13:59] yes [13:59] they announced it yesterday, hold on [14:00] pstolowski: random google https://techcrunch.com/2019/11/13/github-launches-a-mobile-app-smarter-notifications-and-improved-code-search/ [14:00] but there are pics there I didn't see before [14:01] standup [14:01] enjoy [14:01] I'll get breakfast now [14:04] any idea if gnomne-contacts used to be a classic snap? [14:05] mborzecki: is it about that bug that was reported recently? [14:05] zyga: https://bugs.launchpad.net/snapd/+bug/1852355 [14:05] mborzecki: I'll try to find out, I encouraged the reporter to file it [14:05] Bug #1852355: Available snap not refreshed to ("cannot update "gnome-contacts": snap "gnome-contacts" is not a classic confined snap") [14:07] i recall we did some back and forth changes around that are [14:08] yeah, perhaps something in the state was set so that we would never upgrade [14:08] but [14:08] but [14:08] it was only during automatic refresh [14:08] a manual refresh succeeded [14:08] so it's a bug for sure [14:19] duh, we're missing some mocking in gadget remodel unit tests, and spend time waiting for serial to be requested :/ [14:19] added mocking and for instance TestRemodelSwitchGadgetTrack execution time went from 1.3s down to 0.083s [14:23] zyga that new github app looks really cool [14:23] I'm excited for the new notifications stuff though [14:23] that will be nice to filter it more [14:27] mborzecki: re your comments on #7431, the reason I went with `st.Unlock(); defer st.Lock()`, etc. is because Samuele showed me that during a different PR, and actually I am starting to like that kind of code flow because it means when someone else goes to refactor this they don't need to be as careful about returning when st is unlocked because they can return and state is still ok [14:27] PR #7431: overlord/snapstate: don't re-enable and start disabled services on refresh, etc [14:27] but if you like I can still refactor to use your suggestion [14:30] ijohnson: no, it's ok, don't like much that defers keep piling and we end up locking/unlocking (potentially a number of times) inside the function epilogue, but i suppose it's not a big deal [14:31] mborzecki: yes I think it does lead to a bit of thrashing after the function returns but it reads cleanly [14:31] ijohnson: right, let's got with easier to read code for now ;) [14:32] mborzecki: https://xkcd.com/1691/ [14:32] hahah [14:54] * ijohnson relocates back home === ricab|lunch is now known as ricab [15:05] ogra: sorry, other things came in between. I'd have prefered to be able to run docker without sudo, it's not a big thing. [15:07] * cachio lunch [15:14] pokk, well, you can use the --extrausers option to adduser and friends, that creates additional groups and users in a passwd db under /var/lib/extrausers ... [15:15] ogra: but I take it you'd argue it's a very bad idea? [15:15] (which is fine for development etc ... just not really something i'D encourage in a production system for an IoT device) [15:16] it is fine if you use the install as build machine etc ... and typically use ssh logins and the like ... [15:16] ogra: no no :) that I definitly can agree with. This is mainly me playing around with my hmoelab proxmox/ubuntu core [15:16] most IoT gateways, routers, APs and digital signage kiosks wont actually need a user though [15:16] seemed like core would be a good small vm to run docker in, so I'm trying that out [15:17] while docker will surely work, i'd recommend lxd instead ... thats way better integrated (and you can run snaps inside the containers ... thats tricky with docker) [15:19] lxd doesn't seem ideal for running many smaller things. I'll move things like dbs to lxd tho [15:54] cachio: any idea what that means: Nov 14 13:35:11 arch google_clock_skew_daemon[615]: pkg_resources.DistributionNotFound: The 'google-compute-engine==20190801.0' distribution was not found and is required by the application ? [15:55] cachio: just got such spread test failure [15:56] cachio: full log: https://api.travis-ci.org/v3/job/611861148/log.txt [16:25] pstolowski, checking [16:26] pstolowski, seems to be a problem with the package google-compute-engine [16:27] pstolowski, I think it was going to be updated [16:27] pstolowski, I am running that to see if I can fix it [16:28] I'll tell you once I have more info [16:28] cachio: thanks! [16:29] pstolowski, np [16:40] not a happy week for spread tests... [16:41] pstolowski, hehehee [16:43] pstolowski is really weird the error, I'll try to update the image to see if that fix the issue === charles_ is now known as charles === pstolowski is now known as pstolowski|afk [17:16] pstolowski|afk, I tried to update arch [17:16] but I am getting other problems with apparmor package [17:18] pstolowski|afk, images are having a bad week too [20:18] PR snapcraft#2798 opened: cli: improve the remote-build upload messaging [22:42] PR snapcraft#2791 closed: introduce bind-ssh support [22:45] PR snapcraft#2798 closed: cli: improve the remote-build upload messaging [23:06] PR snapcraft#2799 opened: Plainbox test updates