[01:57] PR snapcraft#3887 opened: store: support status for an onprem store [06:53] PR snapd#12059 closed: packaging/*/tests/integrationtests: reload ssh.service, not sshd.service [06:58] PR snapd#12058 closed: overlord: allow seeding in the case of classic with modes system [07:43] PR snapd#12079 opened: multiple: move arguments for auth.NewUser into a struct (auto-removal 1/n) [07:44] Hi [07:44] mardy, testing is snapd.socket is ready is not enough [07:44] https://0x0.st/o90z.txt [07:45] https://0x0.st/o90i.txt [07:53] I can install a snap package in a chroot during os deployment [07:53] so i differ it at startup [07:53] and now i can't even install a pacakge at startup [07:53] guys [07:53] seriously ?? [07:53] s/can/can't [07:54] my playbook is 500 tasks [07:54] it plays with a lot of packages, configurations, daemons [07:54] no other tools is problematic [07:55] the only problem is snap [07:55] snap and snap again [08:13] PR snapd#12080 opened: overlord: track security profiles for non-active snaps [08:18] if i test on snapd.service instead of socket, it seems to wait without end [08:19] ok, does snapd depend on graphical.target ? [08:19] it would be very funny [08:24] mardy i run ansible-pull before lightdm service [08:25] it seems that snapd service is never ready before lightdm [08:25] i can't identify the constraint [08:25] but if i pkill ansible-pull then lightdm shows up and snapd becomes instantly ready [08:25] probleme is that ansible-pull needs snapd to be ready to be able to install packages [08:30] lets move firefox install to flatpak too. Then i remove snapd, that shit is simply not working [08:33] eoli3n, you might hit more problem with flatpak [08:33] i don't have any problem with any flatpak [08:33] i install many [08:34] I mean the firefox one [08:34] maybe, but at least i can install it [08:34] automate the deployment* [08:34] I have not followed what issues you have with it [08:35] just warning you that you might end up with bugs that are fixed on snap [08:35] huhu [08:35] i hear it as a troll [08:36] ? [08:36] what kind of problem do you talk about for exemple ? [08:36] I have no example [08:36] that's what i talk about [08:37] "bugs that are fixed on snap" : no exemple then ? [08:37] because i can tell you much bugs that snap introduce that flatpak doesn't have [08:38] choose your hand [08:38] eoli3n, this is not constructive [08:38] what is not constructive ? [08:38] i explained ealier what are my problem [08:38] s [08:38] I'm actively working on fixing firefox/snap [08:38] I dont have time to also focus on flatpak [08:39] so I am saying it is possible some issues due to running into those environemnt might have been fixed on Snap and not on Flatpak [08:39] saying that much work has been done on firefox snap package without beeing able to give at least one example is not constructive too [08:39] possible yes [08:40] eoli3n https://bugzilla.mozilla.org/showdependencytree.cgi?id=1665641&hide_resolved=0 [08:40] but i can't even deploy firefox package with ansible-pull on my 800 hosts [08:40] I'll let you go through it? [08:40] there are still way too many issues, but so much we can't reproduce and investigate ... [08:40] huhu, my guess is that most of those bugs are introduced by the snapd architecture [08:40] that flatpak doesnt even hit [08:41] maybe [08:41] maybe not [08:41] maybe is the word yes [08:41] there are quite some that are reported on both sides. [08:41] and how many for which we dont know because there's so much less users of flatpak [08:41] i'm not into downing snaps for pleasure [08:41] i would use firefox snap package [08:42] but i'm migrating my infrastructure to ubuntu 22.04, and i need it to be done in next 5h [08:42] then i hit the ultime snapd problem [08:42] i can't install them with ansible-pull [08:42] I was just mentionning that if your problem is deployment, you might want to verify first the level of stability of the flatpak [08:42] maybe you will discover it is more broken than then snap for your usage [08:43] i had the same problem 2 years ago with chromium [08:43] apt install chromium resulted in snap install chromium [08:43] and the same story [08:43] impossible to run chromium with nfs mounted home [08:43] is this unfixable by design ? [08:44] lot of time and work both side with snapcraft and finally moved to flatpak [08:44] never had a problem with it [08:44] but maybe there are more people fixing chromium's flatpak [08:44] the problem is not chromium, it is snap [08:45] https://bugzilla.mozilla.org/showdependencytree.cgi?id=1278719&hide_resolved=1 [08:45] flatpak meta bug [08:45] 224 versus 54 [08:46] notice the hide_resolved= [08:46] https://bugzilla.mozilla.org/showdependencytree.cgi?id=1665641&hide_resolved=1 [08:46] current state is 68 to 54 [08:46] it looks like systemd versus runit for exemple [08:46] snapd is overcomplex, that is the failure [08:47] more complexity had more bugs [08:47] that's subjective [08:48] anyway, if someone shows up here to help me to solve that ansible deployment problem, i would stick with firefox on snap [08:48] as is, i just can't deploy it so... [08:49] eoli3n, two minutes on the list, already three bugs that are just dupes from the snap work as well [08:50] https://bugzilla.mozilla.org/show_bug.cgi?id=1764700 [08:50] reported against flatpak, same root cause as of the snap report, it's not because of snapd [08:50] eoli3n, if you are really in a deadend, why not switching to our binaries? [08:50] either tarball or ppa? [08:51] i didn't know there was a ppa [08:51] that's a better option [08:51] lets check [08:51] thanks [08:52] which one should i use to have the current stable version ? [08:53] ppa , [08:53] ? [08:53] ppa:mozillateam/ppa ? [08:53] yes [08:53] yes I think so [08:54] lets test [09:00] Support for NFS and other remote FS is in the works. The reason why it doesn't work now is that snaps are run by executing a setuid binary, snap-confine, which tries to read the current directory, and that fails on most remote FS (for security reason they don't allow root access) [09:00] we are migrating away from setuid, and use Linux capabilities instead [09:00] but it's not a trivial task [09:02] bug 1973321 [09:02] Bug #1973321: snaps don't start when current working directory is on sshfs [09:02] mardy its working now, i don't know how [09:02] but that's ok [09:02] lissyx the funny thing is, i added the ppa, but as ubuntu wrap firefox package with apt to install with snap, i don't know how to force install from ppa [09:03] mardy, could you help about the snapd at startup problem ? [09:04] not sure i exposed it in the clearest way [09:04] lissyx found a doc about not using the snap package : https://doc.ubuntu-fr.org/firefox#installer_firefox_en_deb_classique_au_lieu_de_snap [09:04] in french, perfect [09:06] eoli3n, is this some enterprise deployment? [09:06] french university [09:06] if so, you might want to make sure you block auto-updates, and this might be complicated with the ppa [09:06] which one? [09:06] Montpellier [09:07] ok it's far enough you wont come to complain :D [09:07] huhu [09:07] you saw that complaining on IRC is in my skills [09:07] ;) [09:08] lissyx how to block auto-updates ? [09:08] pretty sure we have some doc on that but I dont know where [09:08] lissyx even funnier, firefox snap package is defaulty install, so i need to be able to remove it with ansible [09:08] then i hit my snapd not ready problem [09:09] https://support.mozilla.org/en-US/products/firefox-enterprise/deploy-firefox-for-enterprise [09:09] thanks ! [09:09] how do you provision your ubuntu base system? [09:14] mardy, so for speech-dispatcher socket, where should I add it? [09:14] boot pxe on a nfsroot with the iso installer extracted, preseed install, chroot -> ansible-pull, then reboot, ansible-pull again with apps installation enables [09:14] enabled [09:14] desktop.go ? [09:14] well during the preseed you can probably uninstall firefox snap? [09:14] i don't think so, that's my i split my ansible-pull runs : snapd is not running in chroot, because systemd is not running in chroot [09:16] s/my/why [09:17] mardy, ok seb128 reminded me about at-spi one, I'm going to do it next to it [09:21] ah !!! [09:21] https://0x0.st/o90v.txt [09:21] why waiting for multi-user.target !! [09:22] lets disable this [09:23] and then i found this !! in my playbook [09:23] https://0x0.st/o90w.txt [09:23] facepalm [09:23] PR snapd#12081 opened: Bug 1787245 - Grant access to speech-dispatcher socket [09:23] that's something we put with mardy to solve my firefox issue [09:25] lets remove this [09:26] ok, that fixed the nfs home autodetection, if i remove it snapd will not detect that autofs use nfs for home [09:33] PR snapd#12065 closed: store/tooling: support using snapcraft v7+ base64-encoded auth data

=== LetoTheII is now known as LetoThe2nd [12:39] PR snapd#12082 opened: tests: remove NESTED_IMAGE_ID from nested manual tests [12:43] PR snapcraft#3887 closed: store: support status for an onprem store [13:12] lissyx this problem will never end [13:12] the ppa kicked us [13:12] kicked us ? [13:13] as i automate many host that add it in the same time [13:13] it timeouts now [13:13] so i think my public subnet was banned or something [13:13] just use the tarball and follow the doc to disable autoupdate ? [13:13] or it's just infra failure, it happens [13:14] i don't need to disable autoupdate [13:14] i use unnatended upgrade so [13:14] I'm referring to firefox' [13:16] yes, but as said, i use unattended upgrade for firefox package on the ppa [13:16] i want it to be as up to date as possible [13:16] you just said PPA is blocked now [13:17] it timed out during the install process, replaying it on only one host worked [13:17] maybe that's a ddos protection or something [13:17] so I'm just suggesting to move to the tarball we distribute instead of usign the ppa [13:18] and doing so, you might want to disable autoupdate in the firefox install [13:18] i don't want static installations [13:23] eoli3n, static? [13:28] PR snapcraft#3884 closed: cli: parse http-proxy and https-proxy [13:55] lissyx extracting a tarball is a static install, you can't upgrade it without manual intervention [14:01] doesnt the FF tarball install update itself ? [14:01] (via a built in mechanism in FF) [14:03] FF tarball does update itself [14:04] yeah, thought so ... [14:06] I have one place where it's extracted to ~/firefox, and I start it manually from there, and every now and then the shell CURDIR that is in that directory gets confused, because it was deleted an recreated [14:06] I have to cd -; cd - [14:06] then I know ff updated itself [14:08] well, i think it shows a popup too that you should restart (after it did the update) [14:09] * ogra is just gussing ... i do indeed use the snap [14:09] *guessing [14:09] but i used to use the tarball some years ago and that was what it did [14:40] PR snapd#12077 closed: tests: add extra space to ubuntu bionic [17:23] PR snapcraft#3875 closed: providers: delete instances via Executor object [17:38] PR snapcraft#3888 opened: Hotfix/7.1.2 merge [17:56] PR snapd#12082 closed: tests: remove NESTED_IMAGE_ID from nested manual tests [18:48] PR snapcraft#3873 closed: cli: add backwards compatible log enablement for LP [18:53] PR snapcraft#3889 opened: cli: add legacy envvar for login and check stdin [20:23] PR snapcraft#3882 closed: Write passthrough to meta data and add some missing fields [20:23] PR snapcraft#3888 closed: Hotfix/7.1.2 merge [20:48] PR snapcraft#3867 closed: DT-500 Fix Cups for Gtk support in Gnome-42 [21:03] PR snapcraft#3886 closed: ua: specify and enable ua services [21:16] PR pc-amd64-gadget#68 closed: Increase boot partition size for classic gadget [22:23] PR snapcraft#3889 closed: cli: add legacy envvar for login and check stdin [22:33] PR snapcraft#3890 opened: ua: enable ua services in legacy