[00:13] PR snapd#8254 opened: add zoommtg url support [03:25] PR snapd#8255 opened: cmd/snap: make the portal-info command search for the network-status interface [07:34] o/ [07:34] I'm off today [07:34] just came to do paperwork [07:34] push a small test to a PR [07:34] and then disappear in a puff of smoke [07:34] trying to organize time for kids when school is cancelled [08:03] morning [08:13] pstolowski: good morning [08:28] mvo: I pushed an extra test to https://github.com/snapcore/snapd/pull/8241/files [08:28] and I'll file the HR stuff now [08:28] PR #8241: interfaces: work around apparmor_parser slowness affecting uio

[08:31] mvo: done [08:34] pedronis: thank you for reviewing env changes again [09:50] can someone please give a +1 on 8250 ? it's a straight cherry pick with tiny [09:50] conflict [09:50] resolved [09:54] PR snapd#7977 closed: snap: add (hidden) `snap download --indirect` option to download via snapd [09:55] PR snapd#8197 closed: snap: refactor code in `snap download` to prepare for snap downloads via snapd [09:58] PR snapd#8250 closed: tests: mock prune ticker in overlord tests to reduce wait times (2.44) [10:16] hi! Has anyone had a look at https://bugs.launchpad.net/snapd/+bug/1866855 ? Debian + nvidia is broken, unfortunately. [10:16] Bug #1866855: nvidia driver integration is incompatible with Debian [12:05] PR snapd#8256 opened: tests: disable ubuntu-16.04-32:tests/main/lxd [12:22] ppd: I can look next week. Thank you for pointing this out [12:25] zyga: Thanks a lot. I remember there were some vague plans to ship driver libs in snaps, which could simplify the nvidia case. Or am I remembering things incorrectly? [12:28] stgraber: did anything change recently in lxd? we see failures in our GCE lxd 16.04-32 test, inside the containter a "apt update" hangs. it tries to connect to an ipv4 address. the host system is using ipv6 to connect to archive.u.c it seems. any hints for me? [12:28] What version of lxd is that? [12:29] stgraber: 3.22 but actually it could be something else, it seems like inside the container it wants to connect to archive.u.c but outside it connects to the gce apt mirror [12:30] Can you check 'lxc info', and look for the key firewall in there, see if it says xtables or nftables [12:31] We added the right bits for nft to the snap last night so maybe there's something wrong going on with that somehow [12:31] stgraber: says nftables [12:32] Okay, what's the Ubuntu version used for that instance? [12:32] stgraber: 16.04-32 [12:33] stgraber: fwiw, no connectivity inside the container it seems [12:34] stgraber: fwiw this is what we do https://github.com/snapcore/snapd/blob/master/tests/main/lxd/task.yaml#L51 [12:34] Okay, should hopefully be easy to reproduce here [12:34] Works fine on other releases? [12:34] stgraber: yeah, afaict, also fine on 16.04-64 [12:35] stgraber: I keep the gce instance open if you have any questions (but will soon have to leave for lunch) [12:35] stgraber: available via tg for ugent cases though [12:35] Hmm, 32 falling but 64 working is very very weird in this case :) [12:36] mvo, hey [12:38] mvo: can you install nftables and run 'nft list ruleset' on both. Also, the output of 'iptables -L' and 'iptables -t nat -L' would be useful [12:38] I am trying to get the shell by using systemd.debug-shell=1 in run mode when using a vm in my machine and I can't get the shell [12:38] cachio: are you using the version with tpm enabled? [12:39] cachio: if so, the command line is measured and it won't work if you add extra parameters [12:39] cmatsuoka, yes, command line: console=ttyS0 console=tty1 panic=-1 systemd.gpt_auto=0 init=/sbin/init snapd_recovery_mode=run systemd.debug-shell=1 [12:40] cmatsuoka, ahh [12:41] cachio: you can change this by adding the parameter to the command line used when sealing the keyfile, let me see where it is in the code... [12:42] cachio: https://github.com/cmatsuoka/snapd/blob/frankfurt-preview-tpm/cmd/snap-bootstrap/bootstrap/tpm.go#L35 [12:47] cmatsuoka, thanks [12:50] mvo: oh and /var/snap/lxd/common/lxd/logs/lxd.log on both systems too [12:50] mvo: well for this one, just the failing one may be enough [12:50] cachio: yaw, let me know if that doesn't work [14:04] stgraber: sorry, was at lunch and then distracted, will try to get you logs as quickly as I can [14:06] mvo: I think I may have it reproduced on a system here, not 100% sure though [14:07] stgraber: ok, it takes a bit of this test to timeout so just keep me updated if you don't need it anymore [14:35] stgraber: http://paste.ubuntu.com/p/SkPXRyVYYP/ [14:41] hi, has anyone seen an error like this: update.go:85: cannot change mount namespace according to change unmount (tmpfs /snap/maas/4977 tmpfs x-snapd.synthetic,x-snapd.needed-by=/snap/maas/4977/maas-cli/lib,mode=0755,uid=0,gid=0,x-snapd.detach 0 0): device or resource busy [14:41] [14:49] ackk: I haven't, it's probably a zyga question but he is not around today unfortunately [14:49] mvo, oh ok. I've seen it happen a few times, always when refreshing the maas snap. it seems something to do with content interfaces from maas-cli [14:52] ackk: ok [14:54] mvo, I'll file a bug if it happens again with details [14:54] thanks [14:56] ta [14:57] mvo: https://github.com/lxc/lxd/pull/7014 [14:57] PR lxc/lxd#7014: Fix nftables issues on older kernels [15:04] ackk: hey [15:04] zyga, hi [15:04] ackk: it's a thing I understand, I just need reliable instructions to reproduce [15:04] I know about the fact your workflow bumps into that often [15:04] just not sure how to reproduce and help you [15:04] I'm off today and tomorrow [15:04] just visited the office because it's raining and I wanted to lock the window [15:04] and it's +16C outside, what a weird weather [15:05] zyga, well, currently I just had "snap install maas --channel=2.7/edge", then realized I got the wrong one, and "snap refresh --edge" [15:05] *snap refresh --edge maas [15:05] zyga, I've seen it multple times in development too with "snap try this/" then "snap try that/" [15:06] zyga, "maas" has a content interface with "maas-cli" which gets installed/connected automatically [15:06] ackk: if you discard can you reproduce that with the refresh sequence? [15:06] zyga, discard? [15:06] ackk: sudo /usr/lib/snapd/snap-discard-ns maas [15:06] oh [15:06] it's essentially a wipe of the mount namespace [15:07] zyga, you mean discard then refresh? [15:07] zyga, I'll try it next time it happens, let you know [15:09] ackk: discard lets you 'start over' [15:09] if you can get a reproducer from clean slate it effectively gives you [15:09] that's golden [15:09] and that might imply it gets fixed quickly [15:11] zyga, I'll try [15:14] ackk: if you reproduce it I can look, I have two days off of roadmap items :) [15:14] ackk: and other priority bugfixes [15:15] zyga, thanks. I've seen it happen realtively frequently, I should be able to repro if I hit it hard enough :) [15:15] ackk: it's okay if it requires a loop over something [15:15] but I doubt it's that because usually this kind of issue is not a race [15:15] unless the app is racing with something but that's odd because we freeze apps for mount ops [15:15] zyga, this was pretty much a clean container with already-installed core18 and snapd snaps [15:16] zyga, installed maas snap, refreshed from different channel -> error [15:16] it's just not something that happens all the times [15:26] mvo: is master red today? [15:26] zyga: very [15:26] uh? [15:26] what happened? [15:26] zyga: there is a PR to fix this in 8256 [15:27] zyga: but it takes some iterations so far to get it out of the reds [15:28] not MATCH [15:28] zyga: yes [15:28] oh well [15:28] I hope it goes green [15:28] and good luck, not a fun day [15:36] zyga: yeah, I got the wrong team from my tea shop, can't get worse than that [16:21] #8242 is ready for further review, sadly some of the adjustments and extra tests I added made larger again [16:21] PR #8242: many: improve environment handling, fixing duplicate entries

[16:21] *made it [17:31] zyga, https://paste.ubuntu.com/p/FjyMT3Jnkb/ [17:31] (at first attempt) [17:31] zyga, this is a freshly created bionic container [17:39] PR snapcraft#2970 opened: providers: match build provider flag keys to envvar [17:51] PR snapcraft#2968 closed: git: always fetch specified source-commit before using [18:04] zyga, FTR: https://bugs.launchpad.net/snapd/+bug/1867193 [18:04] Bug #1867193: failure refreshing snap with content interface [18:51] ackk: thanks [18:51] ackk: I'll look but tomorrow [18:53] zyga, ty [19:59] PR snapd#8256 closed: tests: fix/improve failing spread tests [20:01] thanks cmatsuoka ! [20:04] PR snapd#8257 opened: tests: backport master test fixes [20:07] mvo: so it was a race then [20:07] cmatsuoka: I think so [20:07] * cmatsuoka hates races [20:07] cmatsuoka: looks like the systemd in sid/arch is a bit slower giving journal data [23:12] PR snapd#8258 opened: interfaces/kubernetes-support: allow autobind to journald socket [23:14] PR snapd#8259 opened: interfaces/kubernetes-support: allow autobind to journald socket - 2.44 [23:31] PR snapcraft#2873 closed: candidate testing [23:34] PR snapcraft#2971 opened: tests: use candidate for autopkgtests