[05:01] morning [05:34] Hey mborzecki [05:35] zyga: hey, noted, i'll investigate reconnect issue in a bit [05:36] pstolowski|afk: hey [05:37] pstolowski|afk: something happened? you're earlier than usual :) [05:39] mborzecki: im not working yet, need to go to school etc and will start as usual. just responded to zyga's issue from the weekend [05:45] Hey [05:45] pstolowski|afk: I can still reproduce it. I can get a coffee and try to help [05:45] zyga: hey [05:47] zyga: thanks, ill be on it in ~1h. my early suspicion is something with conflict checking and retry [05:47] Ok [05:48] Ping me please [06:00] zyga: in the meantime can you send me your stat.json? [06:01] Sure, I saved a copy [06:04] woah [06:05] that state file is 34MB long [06:05] no wonder it takes forever to do anything [06:05] how big are your state.json files? [06:08] Not at my PC right now.. but a few MBs at most afair [06:15] weren't we supposed to garbage collect entries in the state file? [06:21] mborzecki: I think we do but not frequently enough for this [06:34] Hi everyone. I would like a manual review for this: https://dashboard.snapcraft.io/snaps/microk8s/revisions/45/ [06:38] Hi, is there anyone from the snapstore online ? [06:53] jdstrand , anyone the review is urgent, we need it for ODS [06:55] PR snapd#5179 opened: daemon: fix unit tests on arch [06:55] trivial PR ^^ === pstolowski|afk is now known as pstolowski [07:02] re [07:08] Is there any way to snap a jar file. I searched lot, but I don't get anything to create snapcraft.yaml. please give me some examples. [07:17] zyga: are you back from vacation? [07:17] Jack_: https://forum.snapcraft.io/t/how-can-i-snap-a-jar/3306 [07:17] mborzecki: nope [07:17] mborzecki: just not feeling very well today, it's pretty cold in spain [07:18] mborzecki: my son is no longer sick but now I feel affected :/ [07:18] zyga: well, nice & sunny here :) if my phone was less of a potato i'd snap a nice picture [07:18] mborzecki: 14C so far [07:18] oh, my, i think this huge state file is going to kill my emacs :> [07:19] pstolowski: disable font-lock-mode or use fundamental-mode [07:19] mborzecki: thanks i'll try if I regain control ;) [07:19] pstolowski: ^G^G^G [07:20] try sublime :) [07:20] that file has 36M of bytes on one line [07:20] most editors die on that [07:20] (become unusably inefficient) [07:24] even json_pp is struggling ;) [07:24] @Zyga : I already tried this, but did not work for me, it shows unable to read jarfile. I also checked my path. [07:25] can you respond on the thread there? [07:25] many more people participate on the forum than are present on IRC [07:26] I will do that... [07:30] allright, emacs survived and loded the file, although gave up on formatting, so i've a big blob [07:37] pstolowski: have you tried passing it through jq first? [07:39] mborzecki: great idea! it actually works [07:40] and I was too optimistic about emacs. it's unusable ;/ [07:49] pstolowski: well, hard to admit but there's always vim [07:50] :D [07:50] mborzecki: jq to a txt file, then emacs worked [07:53] popey: hello are you there? [08:04] zyga: how many snaps do you have? [08:04] about a dozen [08:04] zyga: on the problematic box [08:05] 39 [08:05] woops :) [08:05] more than a dozen [08:05] I keep snapd off because otherwise it would drain my battery [08:05] zyga: right [08:06] zyga: so, first observations base on the state file [08:09] * zyga is curious [08:11] zyga: there was a core refresh, and this triggered reconnect for all existing connections. the state is full of hook tasks - 77608 of them - and the core "reconnect" task is in "Doing" state (but it's not obvious if they are all related to this refresh; lots (all?) of them are still in default state = not started) [08:11] woooah [08:11] 77K hook tasks [08:11] the number 77608 looks insane [08:12] not sure what happened, trying to make sense out of it [08:12] this is my normal development box [08:12] I just noticed it burns battery very quickly and fan is crazy [08:19] pstolowski: one thing to consider is if this is in the stable release [08:19] or just in edge [08:19] to asses how servere this is [08:19] zyga: reconnect changes landed a few days ago [08:27] pstolowski: refresh to edge and see if you are affected [08:27] hey Chipaca [08:27] zyga: hi [08:29] zyga: yep, i'll, just trying to draw some more conclusions from your state [08:31] anyone else running edge that is affected? [08:38] zyga: affected by what? [08:39] Chipaca: 77608 hook tasks when refreshing core [08:39] zyga: that's more than 3 [08:39] yes, must be inflation [08:40] * zyga is ill today, sucks to spend a week off in bed :/ [08:40] zyga: ok, i think i found a bug, not sure if it fully explains the issue but is definately a problem; the hooks on reconnect (and autoconnect) has change=nil [08:41] *have [08:41] * zyga wonders how to fix affected systems [08:46] zyga: I just refreshed to edge and this didn't happen. 2.32.9+git733.7f74160~ubuntu16.04.1 [08:46] i've 19 snaps [08:51] zyga: when you first noticed it did you stop/start snapd multiple times? [08:51] pstolowski: I stop it every time it starts [08:51] pstolowski: I aborted the change [08:52] but subsequent refresh does the same thing [08:55] zyga: ok. i wonder if this explains the inflation of hook tasks [08:55] pstolowski: how many tasks would you expect? [08:55] I didn't restart snapd 70K time [08:56] zyga: sure. 4 hooks for every connection [08:57] so yeah, still doesn't add up i suppose [08:57] would probably be a few hundreds total [09:00] damn, i cannot trigger it.. edge -> stable -> edge worked [09:02] pstolowski: my core is at 4706 [09:02] maybe start from there [09:03] hmm. how do i do that? --revision is reject for core apparently [09:03] *rejected [09:03] is it? [09:03] why? [09:04] $ sudo snap refresh core --edge --revision=4706 [09:04] error: cannot refresh "core": Access by specifying a revision is not allowed for this Snap. [09:04] you need to be a developer [09:04] of core :/ [09:05] uh oh [09:06] zyga: is this a big no-no or can I get added to a list somewhere? [09:06] I think mvo can add you [09:07] he's off today [09:07] supposedly [09:13] aha [09:16] interesting: https://paste.ubuntu.com/p/8HQHfqQ9SW/ [09:20] wat [09:20] what happens if i accidentally remove /snap/current? how could I get it back? [09:31] eraserpencil: you could carefully recreate it i suppose, it's a symlink [09:39] eraserpencil: or you can disable/enable the snap [09:45] thanks, symlink works [10:12] Could I ask about instructions kyrofa posted in his blog? I always miss kyrofa because of timezones [10:13] I installed Slack using snap, but it does not auto-start on startup. How to make this happen? [10:13] PR snapd#5180 opened: tests/main/snap-service-timer: account for service timer being in the 'running' state [10:13] trivial PR ^^ [10:24] any suggestions where to start looking into erorrs in #5134? [10:24] PR #5134: Shrink image generated with snap prepare [10:51] PR snapd#4600 closed: configstate: validate known core.* options [10:52] review board needs a refresh [11:15] PR snapcraft#2142 closed: Release changelog for 2.42.1 [11:18] PR snapd#4369 closed: interfaces/bulitin: add write permission to optical-drive [11:23] cachio: hi, #4416 fails in spread shellchecks [11:23] PR #4416: tests: performance measurements for basic snapd commands [11:23] cachio: i can fix those and push the changes [11:28] mborzecki, ok [11:28] np [11:33] cachio: pushed shellchecks & merged master there [11:34] mborzecki, great, thanks a lot [11:53] * cachio afk === pstolowski is now known as pstolowski|lunch [12:00] off to pick up the kids, bb for standup [12:10] popey: jdstrand: or anyone from the store, I need help! [12:10] sup? [12:11] popey: there is a review request which is urgent on microk8s [12:11] * jdstrand is looking [12:12] kjackal: I'm going to approve it, but you need to use 'organize' or similar to remove the symlink [12:12] package contains external symlinks: bin/iptables-xml lint-snap-v2_external_symlinks [12:14] thank you jdstrand, I am not sure if this can be removed. We need to use the file from the host for now. [12:14] We have to do a proper packaging of docker. I do not know what else are we going to find in the fiture if we do not [12:14] kjackal: r45 is approved. there are a bunch after that are not. r45 still needs to be published to a channel [12:15] jdstrand: will we go through this review process on every build that has external links? [12:15] kjackal: if you want a later version to be approved, please request a manual remove [12:16] thank you [12:18] kjackal: re each time> yes. but your snap is classic, just use /usr/bin/iptables-xml and don't ship the symlink [12:18] remember, with classic snap there is no mount namespace [12:21] kjackal: re docker> that shouldn't be too difficult> just grab the existing packaging, merge into yours and adjust as necessary [12:23] kjackal: please ping me if you need another revision reviewed [12:24] I will thanks jdstrand === pstolowski|lunch is now known as pstolowski [12:56] re [13:02] error: cannot refresh "snapcraft": cannot query the store for updates: got unexpected HTTP status code 503 via POST to "https://api.snapcraft.io/v2/snaps/refresh" [13:02] just this moment [13:03] did someone trip over the power cable? :-p [13:03] https://status.snapcraft.io/ [13:03] looks good in theory [13:03] could be momentary hiccup [13:04] still broken for me [13:04] cachio: standup? [13:11] jdstrand: hopefully last favour https://dashboard.snapcraft.io/snaps/microk8s/revisions/55/ [13:21] kjackal: done [13:21] kjackal: you still need to publish it [13:22] thank you jdstrand, marcoceppi will appreciate your help. Testing this now [13:23] PR snapd#5174 closed: interfaces/default,process-control: miscellaneous signal policy fixes [13:25] \ [13:44] without using Launchpad, what are some ways I could build snaps for other architectures? [13:51] eraserpencil: build.snapcraft.io ? [13:51] (which uses launchpad under the covers) [13:51] Or maybe travis or circleci? [13:55] I didnt go with launchpad cause it's not opensource nor do i want it distributed [13:57] ah okay [13:57] at this time, there are no options [13:57] you can build with snapcraft on a box somewhere, though [13:57] that's basically what b.sc.io does [13:57] well, you can host private repos on launchpad. [13:58] popey: eraserpencil said "without using Launchpad" [13:58] yeah, he said why [13:58] I'm saying he may be wrong about his reasons why [13:58] ah [13:58] but he may not have the privilege to upload code to LP at all [13:58] https://launchpad.net/+tour/join-launchpad#commercial [13:58] even in a "private repo" [13:58] Yeah, just offering optinos [13:58] Pharaoh_Atem, what box are you talking about [13:58] also *options [13:59] eraserpencil: a computer you set up somewhere [13:59] if you use internally something like Jenkins, you can just set up runners that can run snapcraft [13:59] oh [14:00] the other option is to use lxd locally [14:00] lxd doesn't give foreign arches easily, does it? [14:00] you can build i386 from amd64 [14:00] eraserpencil: which architectures do you need? [14:00] I was looking at lxd.. but the guys at #lxcontainers says lxd uses the host's arch [14:00] arm64 [14:01] yeah, arm64 is tricky [14:01] is qemu + kvm a solution? [14:01] that could work, albeit slow [14:04] unless arm64 vps... [14:04] yeah, i think we had some success with an arm64 vps. [14:04] kyle has played with that, but he's not here right now. [14:05] btw, when i run snapcraft, i often notice it runs on one core only, how could I make it run across all cores? [14:06] what do you mean some success? isnt it no success or possible? [14:06] snapcraft itself runs on one core, but if it spawns make, it usually uses the number of processors/cores [14:07] i don't know what level of success Kyle had, he'll be online soon. [14:07] we can ask him [14:07] is kyle kyrofa? [14:07] ya [14:07] ahh, i need a chat with him. I always miss him because of timezones [14:08] could you elaborate on spawning make [14:08] so if your application uses "make" to build, then snapcraft will spawn it with multiple cores [14:09] like plugin: make? [14:09] ya [14:11] okay. Thanks alot [14:13] np [15:00] * ogra_ wonders what eraserpencil means by "not opensource" ... LP surely is opensource since 9 years ... http://blog.launchpad.net/general/launchpad-is-now-open-source [15:03] ogra_: to use launchpad non-private, the code has to be open-source..which mine isnt [15:03] ah *your* code is not opensource ... now i get it [15:08] re [15:08] pstolowski: hey, any updates? [15:14] zyga: i've found one unrelated non-critical bug. all the problematic tasks have nil change id, i looked at the changes in this code from last few days but couldn't find the cause. with current edge I couldn't reproduce it (and I've expected number of hooks and they have correct change ids). i hope to be able to repro when mvo adds me to the core devs [15:14] zyga: in the meantime i'm still looking at the code [15:18] pstolowski: ack, thank you [15:19] I will keep my snapd offline then, I will be back on Thursday to debug in case we don't find the smoking gun [15:49] eraserpencil, any chance you're around? [15:51] kyrofa: hey! [15:51] eraserpencil, we seem to have very little i.e. zero timezone overlap :D [15:52] if you see a post about ROS on the snapcraft forum by me, please ignore it. I think I JUST figure out the mistake [15:52] eraserpencil, the gstreamer one, or another one? [15:52] aye, GMT+8 does not agree with you. [15:53] eraserpencil, yeah glad you know about the forum, that works very well for this type of thing [15:56] how's your experience on arm64 VPS? [15:57] eraserpencil, packet.net works well [15:58] zyga: ping [15:59] niemeyer: yes? [15:59] eraserpencil, I'm curious though: how do you host your code? Private github? Gitlab? [15:59] zyga: Heya.. do you know anything about this message: [15:59] - Setup snap "core" (4571) security profiles (cannot setup apparmor for snap "core": cannot load apparmor profile "snap-update-ns.core": cannot load apparmor profile: exit status 1 [15:59] zyga: That's on opensuse 42.3.. are we missing some dep there? [16:00] hmmm, it looks like we cannot either compile the profile or load it into the kernel [16:00] is this on your machine? [16:00] can you save the profile somewhere [16:00] zyga: Also, are you on holiday still? Haven't seen any notes in the forum or list [16:00] yes, I'm still off, will be back on Thursday [16:01] zyga: That's opensuse, not my machine [16:01] zyga: This is a plain installation from the ground up [16:01] zyga: Brand new machine [16:01] on 42.3 [16:01] * zyga thinks [16:01] on opensuse apparmor is still disabled [16:01] so if we try to compile this it is surprising [16:01] and we definitely don't have the deps [16:01] (apparmor-profiles, apparmor-parser) [16:02] I was working on enabling those but there are still some issues (those with init scripts) [16:02] Yeah, surprising indeed [16:02] which kernel was that running on [16:02] is this a VM or something else? [16:03] 4.4.104-39-default [16:03] zyga: It's a vm on gce [16:04] well, this checks out: https://en.opensuse.org/Features_42.3#Linux_kernel -- 42.3 is on long-term 4.4 kernel [16:04] zyga: I've installed it by hand, but we probably have something to fix on the packaging there [16:04] oh? did you not use our packages? [16:04] zyga: I did exactly what our instructions say [16:05] zyga: Anyway, thanks for the info [16:05] I see, that's odd, I haven't seen that before [16:05] zyga: Go enjoy your holidays.. we can talk more when you're back and relaxed :) [16:05] :-) [16:05] thanks [16:05] btw, did you hear about the issue that affected my sytem? [16:09] private github [16:09] kyrofa: ^^ [16:16] eraserpencil, have you experimented with build.snapcraft.io? If it worked with private github (as a commercial offering), would it solve your problem? [16:22] PR snapd#5179 closed: daemon: fix unit tests on arch [16:22] zyga: yes, i explained to issue in the standup [16:23] kyrofa, I have not, but if it worked then yes [16:27] eraserpencil, we've been mulling that idea over, I'll pass your interest up the chain === pstolowski is now known as pstolowski|afk [16:40] PR snapd#5181 opened: userd: add the "snap" scheme to the whitelist [18:15] mborzecki, hey [18:29] kyrofa, hey, do you know if snapcraft already supports the timer for the daemons? [18:29] kyrofa, I am using version 2.42 [18:32] cachio, the timer? What is the keyword? [18:33] cachio, worst case you can use passthrough if it's new [18:33] Issues while validating None: The 'apps/runner' property does not match the required schema: Additional properties are not allowed ('timer' was unexpected) [18:33] kyrofa, I see this error [18:34] cachio, indeed, I'm unfamiliar with a "timer" property. Was this discussed in the forum somewhere? [18:34] kyrofa, https://forum.snapcraft.io/t/add-support-for-service-timers/1068/10 [18:36] cachio, indeed, we weren't part of that discussion. You'll need to use passthrough [18:38] kyrofa, ok, thanks [18:54] cachio: what's up? [18:58] mborzecki, nothing, problem solved [18:59] mborzecki, it was about how to use the timer from snapcraft.yaml [19:07] aah ok [19:27] I use Swedish in my Xubuntu 18.04. As soon as I install a snap package I get directories in English. I have ~/Dokument (in Swedish) but when installing a snap package I also get ~/Documents (in English). How can I avoid this? [19:41] * cachio afk [20:37] sm0rux: I think it is a known bug, sadly [20:41] mcphail: Thanks for coming back. Sad news, I thought it was a config issue. [20:42] https://forum.snapcraft.io/t/snap-does-not-respect-directory-settings/3274 - looks as if a fix is emerging... [20:51] hi guys, Should be very easy question for you guys. What are the main differences between Ubuntu server 18.04 “live” and “alternative”? Do they have different purposes? [21:02] mcphail: Thanks a zillion for the link! I thought I was stupid who couldn't fix the locale... [21:05] what is the cloud-init?