eoli3nAs said few days ago : 14:40 +R0Ck3R08:38
eoli3n15:44 +mbond08:38
eoli3n16:33 -mbond08:38
eoli3n19:00 abbe_>abbe08:38
eoli3n19:35 +mbond08:38
eoli3nsorry for this, wrong paste :/08:39
eoli3nso as said few days ago about this : https://bugs.launchpad.net/ubuntu/+source/chromium-browser/+bug/188223208:40
ubottuLaunchpad bug 1882232 in chromium-browser (Ubuntu Focal) "Fresh install of chromium-browser during installer-like process, in a chroot fails (when preparing machines prior to first boot)" [Medium,Fix committed]08:40
eoli3nsnap packages can't work with mounted homes08:40
eoli3nwe use nfs homes08:40
eoli3ncannot open path of the current working directory: Permission denied08:41
eoli3ni think zyga said me that he could fix this08:42
zygaeoli3n: o/08:46
zygaI'm a bit indisposed lately08:46
zygaeoli3n: can you please tell me where you mount nfs homes?08:47
eoli3nunder /home08:48
zygadirectly in /home/foo or in a deeper structure?08:48
zygaeoli3n: can you show me how the mount is declared in either /etc/fstab or in a systemd mount unit please08:49
eoli3nzyga its mounted with autofs08:55
eoli3nprodpeda-samba.domain.fr:/home/p00000012366 on /home/p00000012366 type nfs4 (rw,relatime,vers=4.0,rsize=8192,wsize=8192,namlen=255,soft,proto=tcp,timeo=600,retrans=2,sec=sys,clientaddr=,local_lock=none,addr=
zygaeoli3n: can you paste the fstab line as well please08:56
zygathis is what we use for detection08:56
eoli3nthere is no fstab line08:56
eoli3ni use autofs08:56
zygaah, I see08:56
zygaI'm not familiar with autofs, how does it work?08:57
eoli3nbut i don't get what information you need ?08:58
eoli3nautofs mount a directory when it is accessed08:58
zygasnapd has a global flag that indicates if NFS/CIFS workaround should be enabled08:58
zygawe use two data sources for that08:58
zygawe scan /etc/fstab08:58
zygaand we scan actually mounted filesystems08:58
eoli3noh ok i get what you mean08:58
zygaI'm looking for something that indicates that autofs is in use08:59
zygaso the data is defined in /etc/auto.master?08:59
zygacould you please pastebin /proc/self/mountinfo08:59
zygaperhaps there's an automount mount point in place08:59
zygathat would be useful to detect as well08:59
eoli3nzyga: http://ix.io/2qtb09:01
zyga137 29 0:50 / /home rw,relatime shared:87 - autofs /etc/auto.master.d/home rw,fd=7,pgrp=22588,timeout=300,minproto=5,maxproto=5,indirect,pipe_ino=17339909:01
zygathis is interesting09:01
zygaso we can detect that09:01
zygaI think that's enough for me to fix this now09:01
eoli3nnot really09:01
eoli3nbecause filename could change09:01
eoli3nthere is no rules to be sure that autofs manage homes09:01
zygahaving an autofs on /home feels like a good enough flag09:01
zygaand /etc/auto.master.d/home, does that mention nfs?09:02
eoli3nis there anyworkaround to manually indicates to snapd that i use nfs homes ?09:02
zygano, not at present09:02
eoli3nzyga see my first paste line09:02
zygaah, I see now, thank you :)09:02
zygaI'm curious to ask about /net as well09:03
zygathe /net directory is not something that will work with snaps09:03
zygais that a common convention?09:03
eoli3nthat's nfs installed applications09:03
zygaI see09:03
zygaeoli3n: I've updated https://bugs.launchpad.net/snapd/+bug/1784774 and marked https://bugs.launchpad.net/snapd/+bug/1821193 as duplicate09:07
ubottuLaunchpad bug 1784774 in snapd "snapd is not autofs aware and fails with nfs home dir" [Medium,Triaged]09:07
ubottuLaunchpad bug 1784774 in snapd "duplicate for #1821193 snapd is not autofs aware and fails with nfs home dir" [Medium,Triaged]09:07
eoli3nthanks zyga09:08
zygaeoli3n: fixed09:27
zygaeoli3n: except github 500s09:28
zygaeoli3n: I'll open the pull request once github is operational09:30
zygaeoli3n: https://github.com/snapcore/snapd/pull/893610:11
sil2100cpaelzer_: hey! Looking at the removal of php-horde* packages - just making sure, we also want to remove those that migrated to the release pocket already, right?10:38
sil2100Or does it make sense to leave them there as they migrated?10:38
sil2100Since it seems 16 of those migrated successfully10:39
eoli3nzyga: great, thanks a lot11:40
cpaelzer_sil2100: you mean 1868281 I guess12:43
=== cpaelzer_ is now known as cpaelzer
cpaelzeryes I think that includes these12:44
cpaelzerI'll also ask bryce in standup today to be sure12:44
rafaeldtinococpaelzer: just as a fyi... if you're splitting the logical tags in versions because of me (when reviewing your big qemu merges), no need12:48
sil2100cpaelzer: yeah, removing all of those now, we can always re-sync if needed12:51
kanashirohi o/ today I am starting my +1 maint shift and I started updating the "Things to pick" wiki page section, now I am going to check excuses and pick something to work on13:23
sil2100kanashiro: excellent o/13:38
sil2100kanashiro: I'm looking at the r-cran-httr test failures now - I see those regressed in the release pocket actually as well, but I want to check if they're reported upstream or not before I hint them13:39
kanashirosil2100: I am checking puppetdb, it is stuck in proposed for a while and it hasn't migrated in Debian for almost one year. I'll probably file a RM bug against it13:41
Unit193..If you were super ambitious, you could drop the upper limit in puppet-beaker. >_>13:49
kanashiroUnit193: I feel the Debian maintainer is not having too much time to work on the package, I do not want to fix it now and then need to handle other bugs on my own. Better sync it again when it is well maintained14:02
kanashirobtw it never landed in the release pocket14:02
Unit193I only mention as it holds up net-scp14:04
kanashiroUnit193: ah I see, did you try to relax the net-ssh version constraint in puppet-beaker?14:07
Unit193net-scp*, no bug I figure it's what's needed.14:07
kanashiroif not I can give it a shot14:07
Unit193It's easier when all affected packages are in pkg-ruby. :/14:12
kanashiroUnit193: agreed, one interested developer asked for help to work on puppet packages during the last debian ruby team irc meeting14:14
kanashirothey lack manpower apparently14:15
Unit193I mean, according to UDD same for -ruby...14:19
kanashiroUnit193: after relaxing versions of net-s{sh,scp} the test passed, I'll submit a patch to Debian and upload it to Ubuntu14:52
=== zyga-mbp is now known as zyga
=== heather1 is now known as Guest39331

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!