=== cjwatson_ is now known as cjwatson
[06:31] <mborzecki> morning
[07:33] <mup> PR snapd#7755 closed: cmd/snap-failure: fallback to snapd from core, extend tests <Remodel 🚋> <Created by bboozzoo> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/7755>
[07:51] <mup> PR snapd#7769 opened: cmd/snap-failure: passthrough snapd logs, add informational logging <Simple 😃> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/7769>
=== pstolowski|afk is now known as pstolowski
[08:06] <pstolowski> morning
[08:17] <mborzecki> pstolowski: hey
[09:46] <mborzecki> pstolowski: pekkari: can you take a look at https://github.com/snapcore/snapd/pull/7769 ? super simple
[09:46] <mborzecki> pedronis: ^^
[09:46] <mup> PR #7769: cmd/snap-failure: passthrough snapd logs, add informational logging <Simple 😃> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/7769>
[09:46] <pedronis> a 2nd review of #7764 would be great, it's largely new tests
[09:46] <mup> PR #7764: many: test various kinds of overriding for the snapd snap in Core 20 <Created by pedronis> <https://github.com/snapcore/snapd/pull/7764>
[09:47] <mborzecki> pedronis: sure, will do
=== pedronis_ is now known as pedronis
[10:21] <zyga> hello
[10:21] <zyga> mborzecki: I have something that I wanted to show you
[10:22] <mborzecki> zyga: meet?
[10:22] <zyga> sure, hold on a moment
[10:25] <pedronis> mborzecki: I have a question for you in 7762
[10:27] <zyga> mborzecki: https://github.com/zyga/snapd/commits/wip/dont-reuse-mounts and https://github.com/zyga/snapd/commits/fix/lp-1852361 are the branches of interest
[10:28] <mborzecki> pedronis: we have some code that looks up devices based on structure definition, but it might not be exported atm, need to double check
[10:29] <pedronis> mborzecki: ok, thx
[10:30] <zyga> mborzecki: can you join the standup meet please
[10:34] <zyga> https://www.irccloud.com/pastebin/Bc19KaR2/
[10:40] <zyga> mborzecki: google:ubuntu-18.04-64:tests/regression/lp-1852361
[10:47] <mup> PR snapd#7770 opened: testutil, many: make MockCommand() create prefix of absolute paths <Simple 😃> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/7770>
[10:53] <joc> i have a amd64 device running a UC18 image on which i'm trying to install the docker snap, however if fails with:
[10:53] <joc> error: cannot perform the following tasks:
[10:53] <joc> - Mount snap "docker" (423) (snap "docker" assumes unsupported features: snapd2.40 (try to refresh the core snap))
[10:54] <ogra> did you try to refresh the core snap yet ?
[10:55] <joc> according to snap list these are installed:
[10:55] <joc> core           16-2.42.1     8039  stable    canonical✓  core
[10:55] <joc> core18         20191030      1265  stable    canonical✓  base
[10:56] <ogra> and what does "snap version" output (all lines please) ?
[10:57] <joc> hmm
[10:57] <joc> $ snap version
[10:57] <joc> snap    2.36.3
[10:57] <joc> snapd   2.36.3
[10:57] <joc> series  16
[10:57] <joc> kernel  4.15.0-70-generic
[10:57] <ogra> hmm, very interesting
[10:58] <mup> PR snapd#7771 opened: o/hookstate/ctlcmd: snapctl is-connected command <Created by stolowski> <https://github.com/snapcore/snapd/pull/7771>
[10:58] <ogra> if you are on core18 snapd comes from the snapd snap ... try refreshing that
[10:58] <ogra> (though it should totally auto-refresh... not sure how it can be so behind while the core snaps seem to be nicely up to date)
[10:59] <zyga> mborzecki: I pushed the comment changes again
[10:59] <zyga> mborzecki: to both branches
[10:59] <mborzecki> zyga: ok
[10:59] <zyga> mborzecki: I tg'd mvo to know about the situation
[10:59] <zyga> mborzecki: the most important comment is https://github.com/snapcore/snapd/compare/master...zyga:wip/dont-reuse-mounts?expand=1#diff-6cdce42a784f927f710f0f5241a8c68dR418
[11:00] <zyga> mborzecki: I also adjusted https://github.com/snapcore/snapd/compare/master...zyga:wip/dont-reuse-mounts?expand=1#diff-6cdce42a784f927f710f0f5241a8c68dR493 as mentioned during the call
[11:00] <zyga> I'm wrapping up for today
[11:00] <zyga> good luck guys!
[11:00] <mborzecki> zyga: cool, thanks
[11:02] <mborzecki> running spread with both regression tests
[11:03] <joc> snapd did indeed update, thanks for that hint - now indeed i wonder how it managed to be behind given a `sudo snap refresh` was run before
[11:04] <zyga> ondra: ^^^
[11:04] <zyga> ondra: this is the stuff that fixes things blocking you
[11:05] <zyga> ondra: if you take the diff from https://github.com/snapcore/snapd/compare/master...zyga:fix/lp-1852361?expand=1
[11:05] <zyga> ondra: and build snapd with it
[11:05] <zyga> ondra: and confirm it works on your devices
[11:05] <zyga> ondra: that would be a useful data point
[11:05] <zyga> ondra: we need to discuss and properly package that for release (package as in do more code changes where this can safely land behind a feature flag)
[11:06] <zyga> ondra: but the essence is the same
[11:18] <mup> PR snapd#7772 opened: wrappers: write and undo snapd services on core <Remodel 🚋> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/7772>
[11:47] <ondra> zyga on it :)
[11:50]  * pstolowski lunch
[12:13] <mup> PR snapd#7773 opened: cmd/snap-update-ns: fix overlapping, nested writable mimic handling <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/7773>
[12:14] <mborzecki> zyga: pedronis: ^^ opened the branch, added some description to explain the problem, both spread tests are passing, i've looked the mount ns after the tests and things seemed fine
[12:21] <mup> PR snapcraft#2817 opened: snapcraft/plugins: Add gomod plugin <Created by mhilton> <https://github.com/snapcore/snapcraft/pull/2817>
[12:28] <mup> PR snapd#7769 closed: cmd/snap-failure: passthrough snapd logs, add informational logging <Simple 😃> <Created by bboozzoo> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/7769>
[12:34] <mborzecki> pedronis: do you want to have a quick chat about 7769?
[12:35] <pedronis> mborzecki: you mean 7773 ?
[12:36] <mborzecki> pedronis: right, 7773
[12:36] <pedronis> mborzecki: let's have a chat
[12:37] <pedronis> mborzecki: I'm in the standup
=== ricab is now known as ricab|lunch
[13:38] <ackk> hi, with snappy-debug I see the following message from the sna maas snap: https://paste.ubuntu.com/p/X7pKKCqRCc/ is there a way to tell what's that mknod about?
[13:51] <mup> PR snapcraft#2814 closed: Remove gsettings from comment in kde extension <Created by hellsworth> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2814>
=== ricab|lunch is now known as ricab
[13:54] <mup> PR snapcraft#2802 closed: uprev devel python package requirements <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2802>
[13:57] <mup> PR snapcraft#2804 closed: conda plugin: simplify source url/checksum handling <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2804>
[14:09] <mup> PR snapcraft#2808 closed: repo: fix fetch_binary()'s return type for deb repo <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2808>
[14:09] <mborzecki> ondra: did you manage to run zyga's branch?
[14:09] <ondra> mborzecki yep, just did it
[14:09] <ondra> mborzecki still failing
[14:09] <mborzecki> ondra: duh, logs? :)
[14:10] <mborzecki> ondra: are you sure you're using the right snapd?
[14:10] <ondra> mborzecki https://pastebin.canonical.com/p/TBgSJ8QPcN/
[14:11] <ondra> mborzecki this is build I'm using https://code.launchpad.net/~ondrak/+snap/snapd-fix-lp-1852361
[14:11] <mborzecki> ondra: ha, that's a new thing then, before it failed with EBUSY when trying to roll back some mount changes
[14:11] <mborzecki> zyga: ^^
[14:11] <ondra> mborzecki yep, error is different
[14:11] <ondra> mborzecki and i can confirm snap    2.42.2+git1279.g9c608d9
[14:12] <ondra> mborzecki which is commit from that branch I used for build
[14:12] <mup> PR snapcraft#2806 closed: cli: address type errors for mypy uprev  <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2806>
[14:12] <ondra> mborzecki happy to test more fixes :)
[14:19] <mborzecki> ondra: can you set SNAPD_DEBUG in snapd, restart snapd and try to install the snap again?
[14:19] <ondra> mborzecki you mean refresh?
[14:19] <mborzecki> ondra: yes
[14:22] <mborzecki> ondra: you can drop an ovverride in /etc/systemd/system/snap.service.d/local.conf with [Service]\nEnvironment=SNAPD_DEBUG=1\n
[14:23] <ondra> mborzecki it's fine I updated snapd.service file
[14:25] <ondra> mborzecki https://pastebin.canonical.com/p/YKKKKjV32G/
[14:29] <mup> PR snapd#7764 closed: many: test various kinds of overriding for the snapd snap in Core 20 <Created by pedronis> <Merged by pedronis> <https://github.com/snapcore/snapd/pull/7764>
[14:30] <mup> PR snapcraft#2807 closed: meta: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2807>
[14:45] <mborzecki> ondra: interesting, is this the same device that failed the refresh before? was it restarted since then?
[14:57] <ondra> mborzecki yeah same devce
[14:57] <ondra> device
[14:57] <ondra> mborzecki before test I first tested with stable snapd to make sure it's still broken
[14:58] <ondra> mborzecki BTW when I manually stop services I was able to refresh ( without need to reboot)
[14:58] <mborzecki> ondra: ok, can you try this, stop whatever service is there, /usr/lib/snapd/snap-discard-ns <snap-name>; start the service, refresh the snap
[14:59] <mborzecki> ondra: services from that snap ofc
[15:01] <ondra> mborzecki I cannot restart services they are failing with "cannot update snap namespace: permission denied"
[15:02] <mborzecki> ondra: you should be able to stop them
[15:02] <mborzecki> ondra: as in snap stop <snap> or via systemctl
[15:02] <ondra> mborzecki I stopped them and discarded ns
[15:02] <ondra> but I cannot restart them anymore
[15:03] <ondra> mborzecki and I know refresh worked before when services are stopped
[15:03] <ondra> so running refresh with stopped services will not not help much
[15:04] <mborzecki> ondra: i suspected that the previous version of snap-update-ns left the mount ns in some half assed state, so the new one would fail but differently
[15:13] <mborzecki> ondra: can you add SNAPD_DEBUG=1 to the service environment in the unit file? (or via an override)
[15:17]  * cachio lunch
[15:21] <mup> PR snapd#7774 opened: seed: proper support for optional snaps for Core 20 models <Created by pedronis> <https://github.com/snapcore/snapd/pull/7774>
[15:33]  * zyga acks the results from ondra, asked follow up questions in private and resumes off-work activity 
[15:42] <mup> PR snapcraft#2818 opened: project: pivot `info` from ProjectInfo to Snap <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2818>
[16:15] <ackk> hi, does anyone know if it's possible to run sshd in a confined snap?
[16:15] <ackk> I get this error when trying to ssh in: "fatal: permanently_set_uid: was able to restore old [e]gid [preauth]"
[16:53] <ogra> ackk, during build ?
[16:53] <ackk> ogra, no, when I try to ssh
[16:53] <ogra> ah, sorry ... i'm blinf
[16:53] <ogra> *blind
[16:53] <ackk> ogra, I have ssh running (with some LD_PRELOAD) but I can't log in
[16:53] <ogra> i think your prob isnt ssh but login/pam
[16:54] <ogra> when trying to spawn a terminal for you
[16:55] <ogra> have you tried something like "ssh <host> uptime" (that doesnt spawn a shell but directly execs the command)
[16:58] <ogra> theoretically all you shoudl need for snapping ssh is network and network-bind plugs ... but there is more to a full login than just ssh ...
=== pstolowski is now known as pstolowski|afk
[17:09] <cjwatson> Not sure you can coherently say that the problem is login/pam when the error message above comes from sshd
[17:10] <cjwatson> Probably need to LD_PRELOAD somewhat harder to do a better job of mimicking the uid swap (i.e. it needs to be impossible to swap back)
[17:32] <mup> PR snapd#7775 opened: seed: support extra snaps on top of Core 20 dangerous models <Created by pedronis> <https://github.com/snapcore/snapd/pull/7775>
[18:12] <mup> PR snapcraft#2805 closed: plugins: address type errors for mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2805>
[18:12] <mup> PR snapcraft#2811 closed: extractors: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2811>
[18:18] <mup> PR snapcraft#2812 closed: store: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2812>
[18:21] <diddledan> cjwatson: LD_PRELOAD hard. no harder!
[18:32]  * cachio afk
[18:58] <ackk> ogra, yeah I tried a non-interactive command, same error
[18:59] <ackk> cjwatson, ah, I see
[19:03] <ackk> cjwatson, although, why does it try to change uid if I try to login as root?
[19:03] <mup> PR snapd#7776 opened: interfaces: add login-session-observe for who, {fail,last}log and loginctl <Created by jdstrand> <https://github.com/snapcore/snapd/pull/7776>
[19:07] <mup> PR snapcraft#2819 opened: isort: automatic formatting/sorting of imports <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2819>
[19:07]  * ijohnson unbroke all his snaps (including irccloud) yay
[19:12] <mup> PR pc-amd64-gadget#11 closed: Add mmx64.efi (MokManager) to support mokutil <Created by tsunghanliu> <Merged by vorlonofportland> <https://github.com/snapcore/pc-amd64-gadget/pull/11>
[19:22] <mup> PR snapcraft#2813 closed: pluginhandler: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2813>
[19:27] <mup> PR snapd#7777 opened: snap-confine: suppress noisy classic snap file_inherit denials <Created by jdstrand> <https://github.com/snapcore/snapd/pull/7777>
[19:46] <mup> PR snapcraft#2810 closed: build-providers: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2810>
[19:49] <mup> PR snapd#7778 opened: seccomp: also allow 'chown snap_daemon:root' <Created by jdstrand> <https://github.com/snapcore/snapd/pull/7778>
[19:54] <jdstrand> ogra: hey, I have a note to add this: /sys/firmware/devicetree/base/system/linux,revision and /sys/firmware/devicetree/base/soc/ranges read access to hardware-observe (needs investigating; these are binary)
[19:55] <jdstrand> ogra: but reads on /sys/firmware/** is already allowed via hardware-observe back in 78505e088b79e5ee68e01f23ba5d954dac97c520 (2016). do you recall what this is about?
[20:04] <mup> PR snapd#7779 opened: interfaces: misc updates for u2f-devices, browser-support, hardware-observe, et al <Created by jdstrand> <https://github.com/snapcore/snapd/pull/7779>
[20:25] <mup> PR snapcraft#2816 closed: assorted: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2816>
[20:28] <mup> PR snapcraft#2809 closed: project-loader: address errors from mypy uprev <Created by cjp256> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2809>
[20:49] <mup> PR snapcraft#2820 opened: uprev mypy to 0.740! (and address remaining errors) <Created by cjp256> <https://github.com/snapcore/snapcraft/pull/2820>
[21:33] <ogra> jdstrand, various adafruit libs that are shipped with HW sensors try to access it for whatever reason