[01:09] uh it seems you can use the snapd api to switch a snap to the '' channel [05:39] mwhudson: I will tell John === chihchun_afk is now known as chihchun [06:32] morning === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [06:54] hmm 2019-03-28 22:17:21 Error preparing google:ubuntu-18.04-64:tests/main/ : [06:54] core 6673 stable canonical* broken [07:37] mvo: morning [07:39] mvo: master seems broken now [07:39] mborzecki: hey, morning [07:39] mvo: for some reason there are no mount units for the snaps [07:39] mborzecki: what is the error? [07:40] mborzecki: woah [07:40] mvo: https://paste.ubuntu.com/p/3Kqjg7j76W/ [07:42] mborzecki: the core18 snap changed since yesterday [07:42] mvo: snapd log https://paste.ubuntu.com/p/fNmpTNYMFj/ [07:42] http://people.canonical.com/~mvo/core18-changes/html/edge/unknownr828_unknownr834.html [07:43] but nothing that would explain why mount units are missing [07:43] mborzecki: what do you see in ls /etc/systemd/system/snap* ? [07:44] mvo: no mount units, only snapd.{service,socket} overrides [07:44] mborzecki: woah [07:45] mborzecki: according to travis it started with the most recent master merge but that was just enabling suse again [07:45] mborzecki: so that does not make much sense [07:45] mborzecki: is it just ubuntu-18.04? [07:45] mvo: yes, but sa the same on 18.10 [07:45] s/sa/saw/ [07:46] mborzecki: but not older ubuntus or other distros? [07:46] * mvo scratches head [07:46] mvo: come to think of it, i only saw 18.* [07:47] mborzecki: it does not make much sense, does it :) [07:47] mborzecki: the joy of debugging things [07:47] mborzecki: I will try to reproduce and look for clues [07:49] mvo: checked 4 PRs, only 18.04 [07:55] good morning! === chihchun is now known as chihchun_afk [07:56] mvo: hey, I found some odd things last night === chihchun_afk is now known as chihchun [07:56] I came to check on my tests and found that core18 failed [07:57] while it passed (though I ran the test in isolation) earlier that day [07:57] zyga: hey [07:57] mvo: installing hello-world actually works (?!) [07:58] mborzecki: because snapd thinks core is there [07:58] and refreshing core from --edge worked too, thereis a mount unit now and all === pstolowski|afk is now known as pstolowski [08:07] morning [08:09] heh ausearch -m AVC --checkpoint doesn't work on centos :/ [08:10] pstolowski: fusion update today [08:11] mborzecki: very strange, anyway, qemu running here, I'm curious about the results. maybe one of our scripts goes wild? [08:11] mborzecki: in the tests restore/prepare [08:11] pstolowski: I'm just installing it [08:11] mborzecki: but then we did not change those in a while [08:11] pstolowski: word of caution, it restarts session apps [08:11] iirc [08:11] zyga: ah, thanks for hint, indeed. installing as well [08:11] pstolowski: it will restart browser / etc [08:11] careful :) [08:12] zyga: ha.. ok [08:12] https://docs.vmware.com/en/VMware-Fusion/11/rn/fusion-1103-release-notes.html just fixes === cpaelzer__ is now known as cpaelzer [08:23] https://www.bing.com/th?id=OHR.AurovilleIndia_ROW0132522166_1920x1080.jpg <- amazing wallpaper of the day [08:26] lxd from snaps doesn't work on centos :/ [08:45] rm: cannot remove '/var/cache/snapd/aux': Is a directory [08:45] what is aux? [08:45] is that the auxiliary store stuff? [08:46] yeah [08:47] so, public service announcement: I'm off today [08:47] perhaps I should look at why the world is so red today [08:53] zyga: do you want to take a last look at https://github.com/snapcore/snapd/pull/6634 or can i land it? [08:53] PR #6634: snap: add validation of gadget.yaml [08:54] mborzecki: looking [08:54] let me just page through quickly [08:55] ok [08:59] zyga: i see you're wodnering too about the name/label thing in gadget.yaml [08:59] yeah [08:59] I didn't notice it before [09:00] because of go fmt's reshuffling of the whole blog [09:00] block [09:00] heh [09:00] maybe renaming it to FilesystemLabel would be useful [09:01] zyga: though, a follow up mabye? i'd like to avoid the full ci cycle again, given that it's fragile today [09:01] but what is the difference? [09:02] sure [09:19] mborzecki: hi, name is also used in relative offsets, no? [09:20] mborzecki: almose done [09:24] mborzecki: reviewed [09:25] mborzecki: can you help me get https://github.com/snapcore/snapd/pull/6597 landed [09:25] PR #6597: cmd/snap-update-ns: refactor of profile application (1/N) [09:25] it needs a 2nd review [09:26] alterenatively https://github.com/snapcore/snapd/pull/6502 [09:27] PR #6502: dirs,overlord/snapstate: add Soft and Hard refresh checks [09:27] mvo can you review https://github.com/snapcore/snapd/pull/6410 please [09:27] PR #6410: release-tools: add debian-package-builder [09:32] zyga: sure [09:32] thank you [09:32] mborzecki: hm, strange, "spread qemu:ubuntu-18.04-64" works here with master, thats odd [09:33] zyga: I looked at 6583, it fails to build right now but I can take care of this if you want [09:34] looking [09:35] mvo: hmm [09:35] hmm [09:35] odd [09:35] please let me dig at this [09:35] I have a few PRs that need a 2nd review [09:36] those would be good to land instead :) [09:36] pedronis: yes, name is also used for refrring to structures in relative addressing [09:36] PR snapd#6410 closed: release-tools: add debian-package-builder [09:36] \o/ thanks mvo [09:38] mvo: perhaps this one https://github.com/snapcore/snapd/pull/6502 <- it's not that long [09:38] PR #6502: dirs,overlord/snapstate: add Soft and Hard refresh checks [09:39] haha revolut is preparing for no-deal brexit :P [09:41] what are they doing? [09:43] zyga: need to verify your identity again if you previously used a driver's license for verification [09:44] odd [09:44] how is that related? [09:45] zyga: because they'd apparently need to migrade to a seaprate legal entity taht's in the eu (right now it's uk based) [09:45] lol [09:45] I feel sorry for people in the UK [09:45] monty python live show === chihchun is now known as chihchun_afk [10:03] * Chipaca notes he isn't working today [10:11] * mvo hugs Chipaca and shushes him away [10:14] PR snapd#6665 opened: overlord/ifacestate: implement String() method of HotplugDeviceInfo for better logs/messages [10:16] brb, need some food [10:16] mborzecki: I'll review in a m oment [10:17] zyga: can you finish the review of https://github.com/snapcore/snapd/pull/6629 when you have a moment (needs second +1) [10:17] PR #6629: overlord/snapshotstate: helpers for dealing with snapshot expirations (automatic snapshots 2/N) [10:17] ? [10:22] pstolowski: sure === chihchun_afk is now known as chihchun [10:28] pstolowski: looking now [10:28] zyga: ty! [10:34] pstolowski: done [10:35] pstolowski: you resolved https://github.com/snapcore/snapd/pull/6629#discussion_r270354045 without commenting [10:35] PR #6629: overlord/snapshotstate: helpers for dealing with snapshot expirations (automatic snapshots 2/N) [10:35] ah, sorry, wrong link [10:35] https://github.com/snapcore/snapd/pull/6629/files#r268534384 [10:35] this one [10:40] zyga: hmm it shows me entire PR. do you mean the discussion where you proposed having "snapshots" key defined at the global level to avoid typos? [10:40] pstolowski: cutoff vs cut-off [10:41] line 130 of snapshotstate.go [10:44] zyga: yes, sorry about that, i thought i added it to the batch of github commits but didn't. should be fixed now [10:44] zyga: thanks for the review! [10:44] no worries, it's a bikeshed comment so you can always contend that :) [10:59] pstolowski: reviewed https://github.com/snapcore/snapd/pull/6665#pullrequestreview-220472935 [10:59] PR #6665: overlord/ifacestate: implement String() method of HotplugDeviceInfo for better logs/messages [11:00] * zyga watches https://www.theguardian.com/politics/blog/live/2019/mar/29/brexit-debate-latest-developments-live-news-may-at-risk-of-fresh-defeat-as-mps-debate-withdrawal-agreement-for-third-time-live-news [11:00] any more reviews I can do [11:01] please shoot URLs at me [11:08] anyone [11:20] pstolowski: does hotplug help here? https://forum.snapcraft.io/t/snapd-interfaces-dev-tty-0-9-read-write/10645 [11:25] PR snapd#6666 opened: Change file names that are breaking go modules [11:38] zyga: sorry, i had 1-1, will take a look in a moment [11:38] sure === chihchun is now known as chihchun_afk [11:41] hm trying to track down one selinux denial, idk why it happens, but snap is poking /proc//stat of some process [11:41] i have a hunch that it's hooks and go runtime related [11:42] what does the denial look like? [11:42] * cachio afk [11:46] zyga: https://paste.ubuntu.com/p/SMP6VzNR6r/ [11:47] oh, from snapd itself [11:47] interesting [11:47] yeah [11:47] my bet is go runtime doing something funny [11:47] yes [11:47] what is that pid? [11:48] is 5612 still running [11:48] no [11:48] and it isn't the pid of the shell tha ran snap install|remove [11:48] on top, it's centos, can't perf record what's starting :/ [11:48] maybe it is related to snapctl and peers over sockets somehow [11:48] mborzecki: I'm doing some IO so I cannot search now [11:49] but perhaps check for "proc" and "stat" in golang tree? [11:51] looking gnow [11:54] hmmmm [11:54] switched the policy to non-permissive, and, calling systemctl daemon-reload failed :/ [11:55] zyga: yes, i think we could have a new tty interface that utilizes hotplug and creates single slot for every ttyN [11:55] mborzecki: funn [11:55] mborzecki: I cannot find anything in the tree that does this [11:55] perhaps in our vendor tree? [11:56] mborzecki: polkit [11:56] ./polkit/pid_start_time.go: filename := fmt.Sprintf("/proc/%d/stat", pid) [11:56] it must be polkit [11:57] pstolowski: can you update the thread please? [11:57] omg [11:59] zyga: will do [12:17] * pstolowski lunch [12:20] zyga, mborzecki I have a theory about why ubuntu-18.04 is now failing - I suspect we started seeding cloud-sdk [12:20] in the official images [12:20] oh [12:20] can you tell me moree [12:20] so ubuntu 18.04 in the cloud has changed? [12:21] and what does cloud-sdk do? [12:21] interesting [12:21] that's google cloud sdk? [12:32] * zyga afk === chihchun_afk is now known as chihchun [12:55] ok, so disabling dontaudit and disabling permissive mode is actually enlightening [12:56] i can now install remove lxd snap with selinux in enforcing and without permissive flags in our policy === ricab is now known as ricab|lunch [13:16] Wooot [13:16] That is a great milestone mborzecki [13:27] mborzecki: nice \o/ === ricab|lunch is now known as ricab === chihchun is now known as chihchun_afk [14:48] mvo, 18.04 error should be fixed once the new image is generated [14:49] it could take 20/30 minutes [14:54] cachio: yeah, I think I found the issue, the snapd purge does not work [14:54] cachio: because /var/cache/snapd/aux cannot be removed [14:55] cachio: and new images will fix that so yay! [14:56] mvo, I think so, so far all the tests are passing, also I added some checks to validate the imgae [14:56] cachio: yeah [14:56] cachio: thanks [14:57] mvo, yaw === epod is now known as luk3yx [15:17] zyga: one for you: https://github.com/CanonicalLtd/multipass/pull/702#issuecomment-478036255 [15:18] PR CanonicalLtd/multipass#702: [snap] add libssl1.1 from bionic for Qt's consumption [15:18] zyga: I'll trade you for https://forum.snapcraft.io/t/chaining-dependencies/10581 ;) [15:31] * cachio lunch [15:32] Saviq: looking [15:32] nice, thank you :) [15:32] I'm looking at that in relation to another mount namespace behavior [15:32] but I'm off today so not much progress [15:34] slacker ;) [15:43] Saviq: that last year's quota has to burn :) [15:48] mvo, we've encountered a problem with mir-kiosk starting on core18. With ogra's help I've traced it to "vt.handoff=2" in system-boot/cmdline.txt. Could you remove that please? https://github.com/MirServer/mir/issues/779 [15:53] alan_g: could you please add a github issue against https://github.com/snapcore/pc-amd64-gadget ? [15:53] alan_g: and/or provide a patch there :) technically this is the foundations team now (sil2100 is the main contact currently) [15:53] alan_g: but we still like to help [15:54] mvo, sure. (I get into this so seldom I don't know who to ask) [15:55] mvo, thats a pi issue, not amd64 ;) [15:57] mvo, that should be https://github.com/mvo5/pi-gadget/blob/master/configs/cmdline.txt [15:57] I think [15:58] https://github.com/snapcore/pi-gadget [15:58] that one (i think :) ) [15:59] Ah yeah, too many repos, [15:59] yup [16:03] ogra: oh, sorry [16:03] Hm that one doesn't have any issues. [16:04] alan_g: so you can be the first one! [16:04] and wint a price ! [16:04] alan_g: thanks for raising this btw and no worries, we are here to help [16:04] *win [16:04] mvo, as in nowhere to put them [16:05] send a postcard to sil2100 then [16:05] alan_g: oh, let me look [16:05] It has PRs, so I'll do that [16:05] (after tea break) [16:06] alan_g: \o/ and I will investigate why there are no issues enabled [16:09] alan_g: I actually think this was deliberate that they are disabled. LP seems to be the bugtracker. however it also seems like we need to communicate this better. I will sync with lukasz about this [16:38] mvo, ogra FYI https://github.com/snapcore/pi-gadget/pull/12 [16:38] PR pi-gadget#12: Enable mir-kiosk to start on Ubuntu Core 18/RPi 3 [16:40] alan_g: \o/ [16:41] alan_g: thank you [16:41] mvo, np - the hard part was tracking down where. [16:50] PR # closed: core-build#11, core-build#22, core-build#26, core-build#37 [16:51] PR # opened: core-build#11, core-build#22, core-build#26, core-build#37 === pstolowski is now known as pstolowski|afk [20:09] PR snapcraft#2506 closed: schema, tests: add more detail wrt numeric version errors [20:42] * cachio afk [20:54] PR snapcraft#2519 opened: packaging: use local patchelf